Protein NMR Techniques [2 ed.] 9781588292469, 1588292460

Citation preview

METHODS IN MOLECULAR BIOLOGY

TM TM

Volume 278

Protein NMR Techniques SECOND EDITION Edited by

A. Kristina Downing

Protein NMR Techniques

M E T H O D S I N M O L E C U L A R B I O L O G Y™

John M. Walker, SERIES EDITOR 300. 300 Protein Nanotechnology: Protocols, Instrumentation, and Applications, edited by Tuan Vo-Dinh, 2005 299. 299 Amyloid Proteins: Methods and Protocols, edited by Einar M. Sigurdsson, 2005 298. 298 Peptide Synthesis and Application, edited by John Howl, 2005 297. 297 Forensic DNA Typing Protocols, edited by Angel Carracedo, 2005 296. 296 Cell Cycle Protocols, edited by Tim Humphrey and Gavin Brooks, 2005 295. 295 Immunochemical Protocols, Third Edition, edited by Robert Burns, 2005 294. 294 Cell Migration: Developmental Methods and Protocols, edited by Jun-Lin Guan, 2005 293. 293 Laser Capture Microdissection: Methods and Protocols, edited by Graeme I. Murray and Stephanie Curran, 2005 292. 292 DNA Viruses: Methods and Protocols, edited by Paul M. Lieberman, 2005 291. 291 Molecular Toxicology Protocols, edited by Phouthone Keohavong and Stephen G. Grant, 2005 290. 290 Basic Cell Culture, Third Edition, edited by Cheryl D. Helgason and Cindy Miller, 2005 289. 289 Epidermal Cells, Methods and Applications, edited by Kursad Turksen, 2004 288. 288 Oligonucleotide Synthesis, Methods and Applications, edited by Piet Herdewijn, 2004 287. 287 Epigenetics Protocols, edited by Trygve O. Tollefsbol, 2004 286. 286 Transgenic Plants: Methods and Protocols, edited by Leandro Peña, 2004 285. 285 Cell Cycle Control and Dysregulation Protocols: Cyclins, Cyclin-Dependent Kinases, and Other Factors, edited by Antonio Giordano and Gaetano Romano, 2004 284. 284 Signal Transduction Protocols, Second Edition, edited by Robert C. Dickson and Michael D. Mendenhall, 2004 283. 283 Bioconjugation Protocols, edited by Christof M. Niemeyer, 2004 282. 282 Apoptosis Methods and Protocols, edited by Hugh J. M. Brady, 2004 281. 281 Checkpoint Controls and Cancer, Volume 2: Activation and Regulation Protocols, edited by Axel H. Schönthal, 2004 280. 280 Checkpoint Controls and Cancer, Volume 1: Reviews and Model Systems, edited by Axel H. Schönthal, 2004 279. 279 Nitric Oxide Protocols, Second Edition, edited by Aviv Hassid, 2004

278. 278 Protein NMR Techniques, Second Edition, edited by A. Kristina Downing, 2004 277. 277 Trinucleotide Repeat Protocols, edited by Yoshinori Kohwi, 2004 276. 276 Capillary Electrophoresis of Proteins and Peptides, edited by Mark A. Strege and Avinash L. Lagu, 2004 275. 275 Chemoinformatics, edited by Jürgen Bajorath, 2004 274. 274 Photosynthesis Research Protocols, edited by Robert Carpentier, 2004 273. 273 Platelets and Megakaryocytes, Volume 2: Perspectives and Techniques, edited by Jonathan M. Gibbins and Martyn P. MahautSmith, 2004 272. 272 Platelets and Megakaryocytes, Volume 1: Functional Assays, edited by Jonathan M. Gibbins and Martyn P. Mahaut-Smith, 2004 271. 271 B Cell Protocols, edited by Hua Gu and Klaus Rajewsky, 2004 270. 270 Parasite Genomics Protocols, edited by Sara E. Melville, 2004 269. 269 Vaccina Virus and Poxvirology: Methods and Protocols,edited by Stuart N. Isaacs, 2004 268. 268 Public Health Microbiology: Methods and Protocols, edited by John F. T. Spencer and Alicia L. Ragout de Spencer, 2004 267. 267 Recombinant Gene Expression: Reviews and Protocols, Second Edition, edited by Paulina Balbas and Argelia Johnson, 2004 266. 266 Genomics, Proteomics, and Clinical Bacteriology: Methods and Reviews, edited by Neil Woodford and Alan Johnson, 2004 265. 265 RNA Interference, Editing, and Modification: Methods and Protocols, edited by Jonatha M. Gott, 2004 264. 264 Protein Arrays: Methods and Protocols, edited by Eric Fung, 2004 263. 263 Flow Cytometry, Second Edition, edited by Teresa S. Hawley and Robert G. Hawley, 2004 262. 262 Genetic Recombination Protocols, edited by Alan S. Waldman, 2004 261. 261 Protein–Protein Interactions: Methods and Applications, edited by Haian Fu, 2004 260. 260 Mobile Genetic Elements: Protocols and Genomic Applications, edited by Wolfgang J. Miller and Pierre Capy, 2004 259. 259 Receptor Signal Transduction Protocols, Second Edition, edited by Gary B. Willars and R. A. John Challiss, 2004 258. 258 Gene Expression Profiling: Methods and Protocols, edited by Richard A. Shimkets, 2004 257. 257 mRNA Processing and Metabolism: Methods and Protocols, edited by Daniel R. Schoenberg, 2004 256. 256 Bacterial Artifical Chromosomes, Volume 2: Functional Studies, edited by Shaying Zhao

M E T H O D S I N M O L E C U L A R B I O L O G Y™

Protein NMR Techniques Second Edition

Edited by

A. Kristina Downing Department of Biochemistry, University of Oxford, Oxford, UK

© 2004 Humana Press Inc. 999 Riverview Drive, Suite 208 Totowa, New Jersey 07512 www.humanapress.com All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise without written permission from the Publisher. Methods in Molecular BiologyTM is a trademark of The Humana Press Inc. All papers, comments, opinions, conclusions, or recommendations are those of the author(s), and do not necessarily reflect the views of the publisher. This publication is printed on acid-free paper. ∞ ANSI Z39.48-1984 (American Standards Institute) Permanence of Paper for Printed Library Materials. Production Editor: Angela L. Burkey Cover design by Patricia F. Cleary. Cover illustration: Background: Figure 3A from Chapter 19, “Membrane Protein Structure Determination Using Solid-State NMR,” by Anthony Watts, Suzana K. Straus, Stephan L. Grage, Miya Kamihira, Yuen Han Lam, and Xin Zhao. Inset: Figure 5 from Chapter 15, “NMR Studies of Protein–Nucleic Acid Interactions,” by Gabriele Varani, Yu Chen, and Thomas C. Leeper. For additional copies, pricing for bulk purchases, and/or information about other Humana titles, contact Humana at the above address or at any of the following numbers: Tel.: 973-256-1699; Fax: 973-256-8341; E-mail: [email protected]; or visit our Website: www.humanapress.com Photocopy Authorization Policy: Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by Humana Press Inc., provided that the base fee of US $25.00 per copy is paid directly to the Copyright Clearance Center at 222 Rosewood Drive, Danvers, MA 01923. For those organizations that have been granted a photocopy license from the CCC, a separate system of payment has been arranged and is acceptable to Humana Press Inc. The fee code for users of the Transactional Reporting Service is: [1-58829-246-0/04 $25.00]. Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1 e-ISBN: 1-59259-809-9 ISSN:1064-3745 Library of Congress Cataloging-in-Publication Data Protein NMR techniques / edited by A. Kristina Downing. -- 2nd ed. p. ; cm. -- (Methods in molecular biology ; 278) Includes bibliographical references and index. ISBN 1-58829-246-0 (alk. paper) 1. Proteins--Analysis. 2. Nuclear magnetic resonance spectroscopy. I. Downing, A. Kristina (Anna Kristina) II. Series: Methods in molecular biology (Clifton, N.J.) ; v. 278. [DNLM: 1. Proteins--analysis. 2. Nuclear Magnetic Resonance, Biomolecular--methods. W1 ME9616J v.278 2004 / QU 55 P9672 2004] QP551.P3976 2004 572'.6--dc22 2003027535

Preface When I was asked to edit the second edition of Protein NMR Techniques, my first thought was that the time was ripe for a new edition. The past several years have seen a surge in the development of novel methods that are truly revolutionizing our ability to characterize biological macromolecules in terms of speed, accuracy, and size limitations. I was particularly excited at the prospect of making these techniques accessible to all NMR labs and for the opportunity to ask the experts to divulge their hints and tips and to write, practically, about the methods. I commissioned 19 chapters with wide scope for Protein NMR Techniques, and the volume has been organized with numerous themes in mind. Chapters 1 and 2 deal with recombinant protein expression using two organisms, E. coli and P. pastoris, that can produce high yields of isotopically labeled protein at a reasonable cost. Staying with the idea of isotopic labeling, Chapter 3 describes methods for perdeuteration and site-specific protonation and is the first of several chapters in the book that is relevant to studies of higher molecular weight systems. A different, but equally powerful, method that uses molecular biology to “edit” the spectrum of a large molecule using segmental labeling is presented in Chapter 4. Having successfully produced a high molecular weight target for study, the next logical step is data acquisition. Hence, the final chapter on this theme, Chapter 5, describes TROSY methods for structural studies. In Chapters 6–12 of Protein NMR Techniques, the focus shifts to studies of aligned molecules, beginning with Chapter 6, which describes different options for the preparation of an aligned sample. So many labs have contributed to the development of new media that I believe it will be particularly useful to have this information summarized in an easily digestible format. Residual dipolar coupling (RDC) data acquisition and incorporation into structure calculations are presented in an equally straightforward manner. Chapters 6 and 7 make it clear that RDCs, a powerful source of structural data that are complimentary to NOE-derived distance restraints, can be measured and used routinely. An exciting chapter on the use of RDCs to study protein dynamics highlights the range of information accessible using these methods and, also, their particular potential to inform on motions on a timescale that previously has not been accessible. Having opened the discussion on RDCs and relaxation, there is an elegant description of how the methods can be combined to yield insight into the properties of multimodule constructs.

v

vi

Preface

Three chapters round out this section: Chapter 10 discusses the correct interpretation of relaxation data and the dangers of ignoring anisotropy; Chapter 11 introduces TROSY methods for the study of dynamics; and Chapter 12 provides a detailed and fascinating account of new methods for the characterization of intermediate timescale motions. NMR studies often share common methodological features, and, in other features, they necessarily vary. Chapters 13–15, discussing studies of partially folded states and of protein–protein and protein–nucleic acid complexes, are designed to highlight three special areas of interest. A computational section focusing on the automation of both assignment and structure calculation methods follows. Having personally spent much time on these aspects of structure determination, I find these methods particularly exciting because of their potential to significantly expedite the process. Finally, Chapter 19 comprehensively reviews the application of solid state methods to the study of membrane proteins, a particularly important but difficult class of targets. Editing Protein NMR Techniques has been more work than I had expected, but it was also even more rewarding than I anticipated. I hope that you will find it to be a valuable resource in your research.

A. Kristina Downing

Contents Preface .............................................................................................................. v Contributors ..................................................................................................... ix 1 Screening and Optimizing Protein Production in E. coli Lorraine Hewitt and James M. McDonnell ........................................... 1 2 Isotopic Labeling of Recombinant Proteins From the Methylotrophic Yeast Pichia pastoris Andrew R. Pickford and Joanne M. O’Leary ..................................... 17 3 Perdeuteration/Site-Specific Protonation Approaches for High-Molecular-Weight Proteins Stephen Matthews ............................................................................... 35 4 Segmental Isotopic Labeling for Structural Biological Applications of NMR David Cowburn, Alexander Shekhtman, Rong Xu, Jennifer J. Ottesen, and Tom W. Muir ........................................... 47 5 TROSY-Based Correlation and NOE Spectroscopy for NMR Structural Studies of Large Proteins Guang Zhu, Youlin Xia, Donghai Lin, and Xiaolian Gao .................... 57 6 Media for Studies of Partially Aligned States Kieran Fleming and Stephen Matthews .............................................. 79 7 Residual Dipolar Couplings in Protein Structure Determination Eva de Alba and Nico Tjandra ............................................................ 89 8 Projection Angle Restraints for Studying Structure and Dynamics of Biomolecules Christian Griesinger, Wolfgang Peti, Jens Meiler, and Rafael Brüschweiler ................................................................................. 107 9 Characterizing Domain Interfaces by NMR Luke M. Rooney, Sachchidanand, and Jörn M. Werner ................... 123 10 Characterization of the Overall Rotational Diffusion of a Protein From 15N Relaxation Measurements and Hydrodynamic Calculations Jennifer Blake-Hall, Olivier Walker, and David Fushman ............... 139 11 TROSY-Based NMR Experiments for the Study of Macromolecular Dynamics and Hydrogen Bonding Guang Zhu, Youlin Xia, Donghai Lin, and Xiaolian Gao .................. 161

vii

viii

Contents

12 Measurement of Intermediate Exchange Phenomena James G. Kempf and J. Patrick Loria................................................. 185 13 NMR Studies of Partially Folded Molten-Globule States Christina Redfield ............................................................................. 233 14 Structure Determination of Protein Complexes by NMR Daniel Nietlispach, Helen R. Mott, Katherine M. Stott, Peter R. Nielsen, Abarna Thiru, and Ernest D. Laue .................... 255 15 NMR Studies of Protein–Nucleic Acid Interactions Gabriele Varani, Yu Chen, and Thomas C. Leeper ........................... 16 Using NMRView to Visualize and Analyze the NMR Spectra of Macromolecules Bruce A. Johnson .............................................................................. 17 Automated NMR Structure Calculation With CYANA Peter Güntert .................................................................................... 18 NOE Assignment With ARIA 2.0: The Nuts and Bolts Michael Habeck, Wolfgang Rieping, Jens P. Linge, and Michael Nilges ..............................................................................

289

313 353

379

19 Membrane Protein Structure Determination Using Solid-State NMR Anthony Watts, Suzana K. Straus, Stephan L. Grage, Miya Kamihira, Yuen Han Lam, and Xin Zhao ............................. 403 Index ............................................................................................................ 475

Contributors JENNIFER BLAKE-HALL • Department of Chemistry and Biochemistry, University of Maryland, College Park, MD RAFAEL BRÜSCHWEILER • Gustaf H. Carlson School of Chemistry and Biochemistry, Clark University, Worcester, MA YU CHEN • Departments of Biochemistry and Chemistry, University of Washington, Seattle, WA DAVID COWBURN • New York Structural Biology Center, New York, NY EVA DE ALBA • Laboratory of Biophysical Chemistry, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD A. KRISTINA DOWNING • Department of Biochemistry, University of Oxford, Oxford, UK KIERAN FLEMING • Department of Biological Sciences, Imperial College of Science, Technology and Medicine, London, UK DAVID FUSHMAN • Department of Chemistry and Biochemistry, University of Maryland, College Park, MD XIAOLIAN GAO • Department of Chemistry, University of Houston, TX STEPHAN L. GRAGE • Biomembrane Structure Unit, Department of Biochemistry, University of Oxford, Oxford, UK CHRISTIAN GRIESINGER • Max Planck Institute for Biophysical Chemistry, Göttingen, Germany PETER GÜNTERT • RIKEN Genomic Sciences Center, Yokohama, Japan MICHAEL HABECK • Unité de Bio-Informatique Structurale, Institut Pasteur, Paris, France LORRAINE HEWITT • Laboratory of Molecular Biophysics, Department of Biochemistry, University of Oxford, Oxford, UK BRUCE A. JOHNSON • Molecular Systems, Merck Research Labs, Rahway, NJ MIYA KAMIHIRA • Biomembrane Structure Unit, Department of Biochemistry, University of Oxford, Oxford, UK JAMES G. KEMPF • Department of Chemistry, Yale University, New Haven, CT YUEN HAN LAM • Biomembrane Structure Unit, Department of Biochemistry, University of Oxford, Oxford, UK ERNEST D. LAUE • Department of Biochemistry, University of Cambridge, Cambridge, UK THOMAS C. LEEPER • Department of Biochemistry and Department of Chemistry, University of Washington, Seattle, WA

ix

x

Contributors

DONGHAI LIN • Department of Biochemistry, The Hong Kong University of Science and Technology, Hong Kong, SAR, People’s Republic of China JENS P. LINGE • Unité de Bio-Informatique Structurale, Institut Pasteur, Paris, France J. PATRICK LORIA • Department of Chemistry, Yale University, New Haven, CT STEPHEN MATTHEWS • Department of Biological Sciences, Imperial College of Science, Technology and Medicine, London, UK JAMES M. MCDONNELL • Laboratory of Molecular Biophysics, Department of Biochemistry, University of Oxford, Oxford, UK JENS MEILER • Department of Molecular Biochemistry, University of Washington, Seattle, WA HELEN R. MOTT • Department of Biochemistry, University of Cambridge, Cambridge, UK TOM W. MUIR • Laboratory of Synthetic Protein Chemistry, The Rockefeller University, New York, NY PETER R. NIELSEN • Department of Biochemistry, University of Cambridge, Cambridge, UK DANIEL NIETLISPACH • Department of Biochemistry, University of Cambridge, Cambridge, UK MICHAEL NILGES • Unité de Bio-informatique Structurale, Institut Pasteur, Paris, France JOANNE M. O’LEARY • Department of Biochemistry, University of Oxford, Oxford, UK JENNIFER J. OTTESEN • Laboratory of Synthetic Protein Chemistry, The Rockefeller University, New York, NY WOLFGANG PETI • Department of Molecular Biology, The Scripps Research Institute, La Jolla, CA ANDREW R. PICKFORD • Department of Biochemistry, University of Oxford, Oxford, UK CHRISTINA REDFIELD • Oxford Centre for Molecular Sciences, University of Oxford, Oxford, UK WOLFGANG RIEPING • Unité de Bio-Informatique Structurale, Institut Pasteur, Paris, France LUKE M. ROONEY • Department of Biochemistry, University of Oxford, Oxford, UK SACHCHIDANAND • Department of Biochemistry, University of Oxford, Oxford, UK ALEXANDER SHEKHTMAN • New York Structural Biology Center, New York, NY

Contributors

xi

KATHERINE M. STOTT • Department of Biochemistry, University of Cambridge, Cambridge, UK SUZANA K. STRAUS • Department of Chemistry, University of British Columbia, Vancouver, British Columbia, Canada ABARNA THIRU • Department of Biochemistry, University of Cambridge, Cambridge, UK NICO TJANDRA • Laboratory of Biophysical Chemistry, National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, MD GABRIELE VARANI • Department of Biochemistry and Department of Chemistry, University of Washington, Seattle, WA OLIVIER WALKER • Department of Chemistry and Biochemistry, University of Maryland, College Park, MD ANTHONY WATTS • Biomembrane Structure Unit, Department of Biochemistry, University of Oxford, Oxford, UK JÖRN M. WERNER • Department of Biochemistry, University of Oxford, Oxford, UK YOULIN XIA • Department of Chemistry, University of Houston, TX RONG XU • New York Structural Biology Center, New York, NY XIN ZHAO • Biomembrane Structure Unit, Department of Biochemistry, University of Oxford, Oxford, UK GUANG ZHU • Department of Biochemistry, The Hong Kong University of Science and Technology, Hong Kong, SAR, People’s Republic of China

1 Screening and Optimizing Protein Production in E. coli Lorraine Hewitt and James M. McDonnell Summary Significant improvements in the technologies used for protein production have been driven by impending genome-scale proteomics projects. These initiatives have favored Escherichia colibased expression systems, which allow rapid cloning and expression of proteins at low cost. The range of commercially available molecular biology kits, vectors, affinity tags, and host cell lines have increased dramatically in recent years. For the structural biology community, where protein production is often a rate-limiting step, these developments have made the process of producing and purifying large amounts of protein for structural studies simpler and faster. The large-scale automated screening approaches for optimizing protein production employed by structural genomics initiatives can be adapted to a more practical targeted approach appropriate for individual structural biology groups. This chapter describes simple, rapid screening methods for testing optimal vector/host combinations using a 96-well format. Key Words: Structural biology; protein expression; vectors; fusion proteins; affinity tags; restriction endonucleases.

1. Introduction With the recent successes in whole genome sequencing, the next great challenge for the biochemistry community will be efforts toward describing the structural genome, the characterization of the interaction proteome, and studies of functional proteomics. Nuclear magnetic resonance (NMR) spectroscopy could potentially play a significant role in aspects of all of these projects. Structural and functional analysis of gene products often requires substantial quantities of proteins, and production of this material represents the bottleneck for high-throughput structural studies. Although many functional genomics approaches, such as “proteome chip”-based research, will require much smaller amounts of any individual protein, production and purification of this

From: Methods in Molecular Biology, vol. 278: Protein NMR Techniques Edited by: A. K. Downing © Humana Press Inc., Totowa, NJ

1

2

Hewitt and McDonnell

protein will still be the rate-limiting step in these studies. Consequently, there have been significant efforts in recent years to improve the efficiency of protein production. In practice, there are only a few structural biology laboratories that actually apply true high-throughput approaches to solving structures, but clearly many laboratories benefit from recent improvements in the technologies involved in protein production. The requirements for rapid cloning, expression, and purification of gene products has resulted in development of new vectors, cell lines, and purification protocols that are of general benefit to the structural biology community. In a structural genomics initiative, the process of choosing targets may be rather different than for a structural biology group. Whereas the “low-hanging fruit” represents the primary targets for structural genomics (1), the targets of most structural biology laboratories generally are dictated by the functional activity of the proteins. But the overall aims of maximizing the efficiency of the approach and minimizing the personnel time involved in the process remain the same for both disciplines. As an NMR group with a focus on protein structure and interactions, the production of protein is very often a limiting step, and any improvements in this area are important for our research productivity. Traditional approaches to protein production offer many choices to the experimenter. The first decision is which expression system to use. Because of its well-characterized genetics, ease of manipulation, and low cost, Escherichia coli-based expression systems are the most commonly used. If the protein requires posttranslational modification (glycosylation, phosphorylation, etc.) for biological activity, then it may be impossible to produce a functional product in E. coli, but there are other systems, including yeast, insect, mammalian, or cell-free in vitro translation methods, that are capable of these modifications. The requirement for metabolic labeling (2H, 13C, 15N) of proteins for NMR studies makes the microbial systems far more practical and affordable; nevertheless, it is possible, albeit expensive, to produce labeled protein in baculovirus and mammalian expression systems (2,3). The cell-free in vitro translation systems that produce proteins using bacterial extracts continue to show improvements in yields (4,5) and successes in labeling (6), but these methods are not in general use, and their successful application for high-level protein production has been limited to a few groups. In recent years, the methylotrophic yeast Pichia pastoris has become a powerful host for the heterologous expression of proteins; however, at present, laborious cloning protocols limit it as an expression system for rapid screening. A chapter on protein production and isotopic labeling in the yeast system Pichia pastoris is found elsewhere in this book (see Chapter 2). This chapter focuses on E. coli-based protein expression.

Protein Production in E. coli

3

1.1. From Genes to Proteins The first step in this process is obtaining the complementary DNA (cDNA) of the gene of interest. Deriving a cDNA from a gene without introns is a straightforward application of recombinant DNAmethods. Deriving cDNAs from a gene with introns has historically been made more challenging because of the lack of availability of a full-length cDNA within an existing library, or a more laborious preparation of full-length cDNA from messenger RNA (mRNA) (7). However, ongoing efforts to produce nearly complete cDNA libraries (such as the Mammalian Gene Collection (8) should soon simplify this process. An alternative, but more costly, approach is the commercially available whole gene synthesis (9). It is common practice to establish the cDNA in a cloning vector, which is then used as the source for subcloning into expression vectors. The subcloning process ideally should be “cut and paste” in character. For the sake of improved efficiency and reliability this is routinely achieved by establishing a universal set of restriction endonuclease sites for the cloning vector and different expression vectors. Several new commercial systems are moving away from the use of restriction enzymes in the subcloning process altogether—instead employing a universal entry clone and then a series of compatible recombination vectors. The commercial systems Gateway™ from Invitrogen (Carlsbad, CA) and Creator™ from BD Biosciences (Palo Alto, CA) are examples of this approach (10). There is a vast array of available bacterial expression vectors. The standard approach is to use an inducible T7 RNA polymerase promoter to drive protein production. It is often convenient to produce the gene product as a fusion protein, using the fusion partner to increase expression levels, improve solubility, target to defined subcellular locations, or merely to act as affinity tags to facilitate protein purification. Fusion proteins with cleavable glutathione S-transferase (GST) and polyhistidine, at either N- or C-termini, have been the most commonly employed, with the polyhistidine tag being the choice of many large-scale automated structural genomics efforts. This may be particularly attractive to structural studies by NMR, in which the presence of a short flexible terminus generally is innocuous for structural analysis. The vectors can then be transformed into numerous available E. coli host strains. Some of the host strains are good generic lines for protein production, whereas others have more specialized function. For example, the Rosetta™ (Novagen, Madison, WI) and Codon Plus™ (Stratagene, Cedar Creek, TX) lines encode genes for transfer RNA (tRNA) for rare codons, allowing for more efficient production of eukaryotic proteins that contain rarely used E. coli codons. Certain vector/host combinations are optimized for specialized function. Novagen uses a pET-32 vector to make a thioredoxin fusion protein in the host strain Origami™, which provides mutations in thioredoxin reductase (trxB) and

4

Hewitt and McDonnell

Table 1 Standard Expression Vectors Vector

Supplier

pET vector series

Novagen

pCAL vectors pMAL vectors

Stratagene New England Biolabs

Characteristics IPTG-inducible T7 promoter (11). Broad range of vectors giving wide choice of multiple restriction sites and affinity tags. For example: pET32 series has a Trx fusion and hexahistidine tag. pET41 series is a GST fusion system with both hexahistidine and S-tag options. pET43.1 series has a hexahistidine affinity tag and a NusA fusion for solubility. CBP (calmodulin-binding peptide) affinity tag. Maltose Binding Protein affinity tag with protease cleavage site; pMAL-p2x has a signal sequence to direct protein to periplasm, which improves solubility.

Table 2 Standard Expression Hosts Cell line BL21 (DE3) B834 C41 (BL21 derived) BL21 AI

Supplier Various Not commercially available (12) Invitrogen

BL21 Codon Plus™ Stratagene AD494 Novagen Origami™ OrigamB™

Novagen

Rosetta-gami™

Novagen

Characteristics Use with T7 promoter systems. As BL21 but adapted for increased expression of “toxic” proteins. Arabinose induced tighter regulation to improve expression of toxic proteins. Codon usage bias. TrxB mutants to facilitate disulfide bond formation. Trx/gor mutants enhance disulfide bond formation giving high yields of properly folded, active protein. Have both the codon usage plasmid to adjust for codon bias as well as the trxB and gor mutations.

Protein Production in E. coli

5

glutathione reductase (gro) genes, to enhance disulfide formation in the bacterial cytoplasm. A list of standard expression vectors and hosts can be found in Tables 1 and 2. Unfortunately no single expression vector or host line works well for all protein constructs; lack of success is typically because of either insufficient protein production or solubility. To date, it has not been possible to reliably predict the expression characteristics of an untested protein. Therefore, it is advisable to screen a matrix of expression vector/host combinations and culture conditions. It is clear that most of the steps in protein production, as simple liquid handling protocols, have the potential to be handled robotically in an automated process. Steps that do not lend themselves to robotic handling can be replaced, such as using magnetic bead-based separations instead of centrifugation steps. However, liquid-handling robots have yet to replace centrifuges in most laboratories. Thus, in the interest of practical applicability, this chapter focuses on more targeted screening approaches for testing protein expression using equipment routinely available in most structural biology laboratories. 2. Materials 2.1. Polymerase Chain Reaction (PCR) and Agarose Gels The percentage of agarose gel is dependent on the size of the DNA to be analyzed (see Table 3). DNA fragments 99 At %) and 2H,13C-methanol (>99 At %) were obtained from CK Gas Products (Finchampstead, UK). 2.1. Seed Culture for Uniform in Shake Flasks or a Fermenter

13C,15N-Labeling

1. Potassium phosphate solution, pH 6.0 (per liter): Dissolve 23 g potassium dihydrogen phosphate and 118 g dipotassium phosphate in water. If necessary, adjust the pH to 6.0 ± 0.1 with potassium hydroxide (corrosive) or phosphoric acid (85% wt; corrosive) (see Note 1). Autoclave and store at room temperature for up to 1 yr. 2. Yeast nitrogen base (YNB) solution (per liter): Dissolve 34 g YNB without amino acids or ammonium sulfate (Difco, UK) in water. Filter-sterilize and store at 4ºC for up to 1 yr. 3. Biotin solution (per 100 mL): Dissolve 20 mg D-biotin in 10 mM sodium hydroxide. Filter-sterilize and store at 4ºC for up to 3 mo. 4. 13C-glucose solution (per 25 mL): Dissolve 1.25 g U-[13C]-D-glucose in water. Filter-sterilize and store at 4ºC for up to 1 yr (see Notes 2 and 3). 5. 15N-ammonium sulfate solution (per 50 mL): Dissolve 5 g 15N-ammonium sulfate in water. Filter-sterilize and store at 4ºC for up to 1 yr (see Notes 2 and 3). 6. 13C,15N-labeled buffered minimal glucose (13C,15N-BMD), pH 6.0 (per 50 mL): Autoclave 30 mL of water and allow to cool to room temperature. Aseptically, add 5 mL of potassium phosphate, pH 6.0, solution, 5 mL of YNB solution, 0.1 mL of

22

Pickford and O’Leary biotin solution, 5 mL of 13C-glucose solution, and 5 mL of 15N-ammonium sulfate solution. Store at 4ºC for up to 1 mo.

2.2. Uniform 1.

13

13C,15N-Labeling

in Shake Flasks

15

C, N-BMD, pH 6.0 (per 200 mL): Scale-up from the 50-mL recipe in Subheading 2.1., step 6. 2. 13C-methanol feed solution (per 100 mL): Add 10 mL 13C-methanol (toxic) to 90 mL water. Filter-sterilize and store at 4ºC for up to 1 mo (see Notes 2 and 3). 3. 13C,15N-labeled buffered minimal methanol (13C,15N-BMM), pH 6.0 (per liter): Autoclave 600 mL of water and allow to cool to room temperature. Aseptically, add 100 mL of potassium phosphate, pH 6.0, solution, 100 mL of YNB solution, 2 mL of biotin solution, 100 mL of 13C-methanol solution, and 100 mL of 15Nammonium sulfate solution. Store at 4ºC for up to 1 mo.

2.3. Fermenter Vessel Preparation for Uniform

13C,15N-Labeling

A 1-L glass fermentation vessel was obtained from Electrolab (Tewkesbury, UK) with fittings for variable-speed impeller, temperature control via “cold finger” and heater mat, adjustable sterile air supply introduced through a sparger, and an air outlet fitted with a condenser. The vessel head plate has ports for temperature, pH, dissolved oxygen (dO2), and foam probes, for acid-, base-, antifoam-, and nutrient-feed lines, and for inoculation and sampling. The fermenter was operated using a controller (from Electrolab) with feedback regulation of temperature, pH (via a base pump), aeration (via the impeller), and foaming (via an antifoam pump). A manually controlled, variable-speed, nutrient-feed pump (from Electrolab) was used to deliver methanol. 1. Basal salts medium (per 400 mL): Dissolve 0.47 g calcium sulfate, 7.5 g magnesium sulfate heptahydrate, and 13.4 mL orthophosphoric acid (85% wt; corrosive) in water. Adjust the pH to 3.0 with solid potassium hydroxide (corrosive) (see Note 4). Sterilize in situ by autoclaving (see Subheading 3.3., step 3). 2. Pichia trace metals (PTM1) (per 100 mL): Dissolve 0.6 g copper (II) sulfate, 8 mg sodium iodide, 0.3 g manganese sulfate monohydrate, 20 mg sodium molybdate dehydrate, 2 mg boric acid, 50 mg cobalt chloride, 2 g zinc chloride, 6.5 g iron (II) sulfate, 0.5 mL sulfuric acid (98%; corrosive) in water. Store at room temperature for up to 1 yr. Filter-sterilize before use.

2.4. Uniform

13C,15N-Labeling

13C-methanol

in a Fermenter

1. feed solution (per 100 mL): Prepare as in Subheading 2.2., step 2. 2. Antifoam solution: Autoclave polypropylene glycol 1025, and store at room temperature indefinitely (see Note 5).

Isotopic Labeling of Pichia pastoris Proteins

23

3. Base control solution: Prepare and autoclave 2 M potassium hydroxide (corrosive), and store at room temperature indefinitely (see Note 6).

2.5. Seed Culture for Uniform 2H,13C,15N-Labeling in a Fermenter 1. Deuterated potassium phosphate solution, pH 6.0 (per 10 mL): Prepare potassium phosphate solution (see Subheading 2.1., step 1). Lyophilize 10 mL of the solution, and resuspend in 10 mL D2O. Filter-sterilize and store at room temperature for up to 1 yr (see Note 2). 2. Deuterated YNB solution (per 10 mL): Prepare YNB solution (see Subheading 2.1., step 2). Lyophilize 10 mL of the solution, and resuspend in 10 mL D2O. Filter-sterilize and store at room temperature for up to 1 yr (see Note 2). 3. Deuterated biotin solution (per 10 mL): Prepare biotin solution (see Subheading 2.1., step 3). Lyophilize 10 mL of the solution, and resuspend in 10 mL D2O. Filter-sterilize and store at room temperature for up to 1 yr (see Note 2). 4. Deuterated 13C-glucose solution (per 30 mL): Dissolve 1.5 g U-[13C]-glucose in 30 mL D2O. Lyophilize the solution, and resuspend in 30 mL D2O. Filter-sterilize and store at 4ºC for up to 1 yr (see Notes 2 and 3). 5. Deuterated 15N-ammonium sulfate solution (per 55 mL): Dissolve 5.5 g 15Nammonium sulfate in 55 mL D2O. Lyophilize the solution, and resuspend in 55 mL D2O. Filter-sterilize and store at 4ºC for up to 1 yr (see Notes 2 and 3). 6. 70%-deuterated 13C,15N-BMD (70%-2H,13C,15N-BMD), pH 6.0 (per 50 mL): Filter-sterilize 30 mL D2O. Aseptically, add 5 mL of deuterated potassium phosphate, pH 6.0, solution, 5 mL of deuterated YNB solution, 0.1 mL of deuterated biotin solution, 5 mL of deuterated 13C-glucose solution, and 5 mL of deuterated 15N-ammonium sulfate solution. Store at 4ºC for up to 1 mo. 7. 100%-deuterated 13C, 15N-BMD (100%-2H,13C,15N-BMD), pH 6.0 (per 50 mL): Filter-sterilize 30 mL D2O. Aseptically, add 5 mL of deuterated potassium phosphate, pH 6.0, solution, 5 mL of deuterated YNB solution, 0.1 mL of deuterated biotin solution, 5 mL of deuterated 13C-glucose solution, and 5 mL of deuterated 15Nammonium sulfate solution. Store at 4ºC for up to 1 mo.

2.6. Fermenter Vessel Preparation for Uniform 2H,13C,15N-Labeling 1. Deuterated BSM (per 390 mL): Dissolve 27.2 g potassium dihydrogen phosphate and 12.4 g magnesium sulfate heptahydrate in 100 mL D2O. Lyophilize the solution, and resuspend in 390 mL D2O (see Note 4). Filter-sterilize and store at room temperature for up to 1 yr (see Note 2). 2. Deuterated calcium chloride solution (per 10 mL): Dissolve 1.47 g calcium chloride dihydrate in 10 mL D2O. Lyophilize the solution and resuspend in 10 mL D2O. Filter-sterilize and store at room temperature for up to 1 yr (see Note 2). 3. Deuterated PTM1 (per 10 mL): Prepare PTM1 solution (see Subheading 2.3., step 2). Lyophilize 10 mL of the solution, and resuspend in 10 mL D2O. Filter-sterilize and store at room temperature for up to 1 yr (see Note 2).

24

Pickford and O’Leary

2.7. Uniform 2H,13C,15N-Labeling in a Fermenter 1. Deuterated 2H,13C-methanol feed solution (per 100 mL): Add 10 mL 2H,13Cmethanol (toxic) to 90 mL D2O. Filter-sterilize and store at 4ºC for up to 1 mo (see Notes 2 and 3). 2. Antifoam solution: Prepare as in Subheading 2.4., step 2. 3. Deuterated base control solution (100 mL): Dissolve 21.2 g tri-potassium phosphate (corrosive) in 100 mL D2O. Filter-sterilize and store at room temperature for up to 1 yr (see Notes 2 and 6).

2.8. Protein Purification by Cation-Exchange Chromatography SP-Sepharose Fast Flow was supplied by Amersham Biosciences (Little Chalfont, UK). 1. Cation-exchange equilibration buffer (per liter): Prepare 20 mM sodium citrate, 0.02% (w/v) sodium azide (highly toxic), pH 3.0. Store at room temperature for up to 1 yr. 2. Cation-exchange elution buffer (per liter): Prepare 20 mM sodium citrate, 1 M sodium chloride, 0.02% (w/v) sodium azide (highly toxic), pH 3.0. Store at room temperature for up to 1 yr.

2.9. Glycoprotein Purification by Lectin Affinity Chromatography Con A Sepharose 4B was supplied by Amersham Biosciences (Little Chalfont, UK). 1. Con A regeneration buffer: Prepare 20 mM Tris, 0.02% (w/v) sodium azide (highly toxic), pH 8.5. Store at room temperature for up to 1 yr. 2. Con A equilibration buffer: Prepare 20 mM Tris, 0.5 M sodium chloride, 1 mM manganese (II) chloride, 1 mM calcium chloride, 0.02% (w/v) sodium azide (highly toxic), pH 7.4. Store at room temperature for up to 1 yr. 3. Con A elution buffer: Prepare 0.5 M methyl-α-D-glucopyranoside, 20 mM Tris, 0.5 M sodium chloride, 0.02% sodium azide (w/v) (highly toxic), pH 7.4. Store at room temperature for up to 1 yr.

3. Methods 3.1. Seed Culture for Uniform or a Fermenter

13C,15N-Labeling

in Shake Flasks

1. Thaw an aliquot of a P. pastoris strain that has been stored as a 20% (v/v) glycerol stock at −80ºC. Pellet the cells by centrifugation at 3000g for 5 min at room temperature. Resuspend the cell pellet in 1 mL of 13C,15N-BMD. 2. Inoculate a sterile 250-mL baffled flask containing 49 mL of 13C,15N-BMD with the cell suspension and grow, shaking at 30ºC for 18–24 h, until the optical density (OD)600 >2 (see Note 7). This will constitute the seed culture for a 1-L shake-flask growth (see Subheading 3.2., step 1) or 0.5-L fermentation (see Subheading 3.4., step 2 and Note 8).

Isotopic Labeling of Pichia pastoris Proteins 3.2. Uniform

13C,15N-Labeling

25

in Shake Flasks

1. Inoculate a sterile 2-L baffled flask containing 200 mL of 13C,15N-BMD with the 50 mL seed culture (see Subheading 3.1., step 2). Incubate with shaking at 30ºC for 48–60 h, until the OD600 = 10–20 (see Note 7). 2. Pellet the cells in sterile containers at 3000g for 5 min at room temperature. Discard the supernatant. 3. Gently resuspend the cells in 1 L of 13C,15N-BMM. Equally divide the cell suspension into four 2-L baffled flasks (see Note 9). Incubate with shaking at 30ºC, until the optimal induction time is reached (see Note 10). Take samples regularly (every 8–12 h) for analysis of cell density (OD600) and of expression level by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (see Note 11). 4. Every 24 h during the induction phase, feed the cultures by adding 12.5 mL 13Cmethanol feed solution. 5. At the end of the induction phase, pellet the cells by centrifugation at 10,000g for 10 min at 4ºC (see Note 10). Remove the yellow-green supernatant, measure its volume, and store it at 4ºC (see Note 12). Resuspend the cell pellet in a volume of 100 mM HCl equivalent to the original supernatant volume. Pellet the cells by centrifugation, as above; combine this supernatant with the original supernatant, and clarify this solution by passing through a 0.2 µM filter (see Note 13). Start purification immediately to minimize proteolytic degradation (see Subheading 3.8., and see Note 14).

3.3. Fermenter Vessel Preparation for Uniform

13C,15N-Labeling

1. Calibrate the fermenter pH probe with pH 4.0 and pH 7.00 buffers, rinse it with deionized H2O, and introduce it into one of the vessel’s ports (see Note 15). 2. Test the fermenter dO2 probe with zeroing gel (available from Mettler-Toledo, Leicester, UK), rinse it with deionized H2O, and introduce it into one of the vessel’s ports (see Note 16). If available, connect a polarization module to the electrode (see Note 17). 3. Mask the external ends of all ports and lines of the fermenter vessel with aluminum foil and autoclave tape, clamp any lines that will be in contact with the media (e.g., the air inlet and sampling line), and sterilize the vessel containing 400 mL of BSM. Also sterilize three empty aspirators for base control, antifoam, and methanol feed (see Note 5). 4. Assemble the vessel on the fermentation system, insert the temperature probe, and connect the facilities for temperature control (e.g., heating mat and cold finger, or vessel jacket with circulating water bath). Allow the media to equilibrate to 30ºC. 5. Connect the impeller and the dO2 probe to the controller. If a polarization module has not been used, allow 6 h for probe polarization before inoculating the vessel with the starter culture (see Note 17). The dO2 level can now be controlled by feedback regulation of the impeller speed (see Note 18). 6. Aseptically, fill the base control aspirator with base control solution, and connect it to the vessel via the base control pump. Connect the pH probe to the controller.

26

Pickford and O’Leary

The system can now regulate the culture pH (see Note 6). Allow the media to equilibrate to the desired pH for fermentation (see Notes 1, 4, and 19). 7. Aseptically, fill the antifoam aspirator with antifoam solution, and connect it to the vessel via the antifoam pump. Connect the foam probe to the controller to allow the system to manage excessive frothing (see Note 5). 8. Set the impeller speed to 250 rpm to aid temperature equilibration. Aseptically, introduce 20 mL 13C-glucose solution, 45 mL 15N-ammonium sulfate solution, 2.5 mL biotin solution, and 2.5 mL PTM1 via the inoculation port (see Note 19). 9. Connect the sterile air-line, open the air outlet, and increase the impeller speed and air-flow rates to the maximum value to be used during the fermentation, typically 1000 rpm and 5 vol of air per volume of culture per min (vvm). After 10 min, calibrate the dO2 probe at 100% saturation. Set the air-flow rate to 2 vvm.

3.4. Uniform

13C,15N-Labeling

in a Fermenter

1. Activate the aeration feedback control with a target dO2 of 20%, and maximum and minimum impeller speeds of 1000 rpm and 200 rpm, respectively (see Note 18). The saturated media should cause the impeller speed to drop to 200 rpm. 2. Inoculate with the 50-mL seed culture (see Subheading 3.1., step 2), and start the controller to monitor dO2, impeller speed, temperature, and pH. Take a 10-mL sample via the sample port for assay and determination of wet cell weight and/or OD600 (see Note 11). 3. Aseptically, fill the methanol feed aspirator with the 13C-methanol feed solution, and connect the aspirator to the vessel via the nutrient-feed pump. 4. When the dO2 “spikes” (typically 18–24 h after inoculation), take a 10-mL sample for analysis, and add 50 mL 13C-methanol feed solution (toxic) via the inoculation port (see Note 20). Start the methanol feed at 2.5 mL/h, increasing to 5 mL/h after 6 h. Take samples regularly (every 8–12 h) for analysis of cell density (OD600) and of expression level by SDS-PAGE (see Note 11). Continue the fermentation until it has consumed at least 250 mL methanol feed solution (see Note 3). 5. To prevent methanol accumulating in the vessel, but at the same time maintaining optimal cell growth, the feed rate of methanol can be increased with time, but it must remain limiting. Under conditions of limiting carbon source, the respiration rate rises and falls as methanol is constantly depleted, then repleni- shed, in the vessel; this is signaled by multiple dO2 “spikes” (see Note 21). 6. Harvest the expressed protein by centrifugation as described in Subheading 3.2., step 5, and start purification immediately to minimize proteolytic degradation (see Subheading 3.8. and Note 14).

3.5. Seed Culture for Uniform 2H,13C,15N-Labeling in a Fermenter To produce significant quantities of 2H,13C,15N-labeled protein, P. pastoris cells must be “conditioned” for growth in a deuterated environment. This is achieved by subculturing the cells into first 70-, then 100%-deuterated media.

Isotopic Labeling of Pichia pastoris Proteins

27

1. Thaw an aliquot of a P. pastoris strain that has been stored as a 20% (v/v) glycerol stock at −80ºC. Pellet the cells by centrifugation at 3000g for 5 min at room temperature. Resuspend the cell pellet in 1 mL of 70%-2H,13C,15N-BMD. 2. Inoculate a sterile 250-mL baffled flask containing 49 mL of 70%-2H,13C,15NBMD with the cell suspension and grow, shaking at 30ºC for 36–48 h, until the OD600 >2 (see Note 7). Pellet a 10-mL aliquot of the cells by centrifugation at 3000g for 5 min at room temperature. Resuspend the cell pellet in 1 mL of 100%2H,13C,15N-BMD. 3. Inoculate a sterile 250-mL baffled flask containing 49 mL of 100%-2H,13C,15NBMD with the cell suspension and grow, shaking at 30ºC for 36–48 h, until the OD600 >2 (see Note 7). This will constitute the seed culture for a 0.5-L fermentation (see Subheading 3.7., step 2 and Note 8).

3.6. Fermenter Vessel Preparation for Uniform 2H,13C,15N-Labeling To achieve the highest levels of deuteration, every care should be taken not to introduce water into the fermenter vessel. 1. Calibrate the fermenter pH probe with pH 4.0 and pH 7.00 buffers, and immerse its electrode in a measuring cylinder of deionized H2O in preparation for autoclaving (see Note 15). 2. Test the fermenter dO2 probe with zeroing gel, rinse it with deionized H2O, and immerse its electrode in a measuring cylinder of deionized H2O in preparation for autoclaving (see Note 16). If available, connect a polarization module to the electrode (see Note 17). 3. Mask the external ends of all ports and lines of the fermenter vessel with aluminum foil and autoclave tape, and clamp any lines that will be in contact with the media (e.g., the air inlet and sampling line). Cover the ports to be used for the pH and dO2 probes with gauze. Sterilize the vessel, the pH and dO2 probes, and three empty aspirators for deuterated base control, antifoam, and deuterated methanol feed. 4. Dry the sterile fermenter vessel in an oven at approx 80ºC for 24–48 h to remove any traces of water. Once cooled, aseptically fill the dry vessel with 390 mL deuterated BSM and 10 mL of deuterated calcium chloride solution. 5. Aseptically, air-dry the sterile pH and dO2 probes for 10 min, rinse their electrodes with filter-sterilized D2O to remove any traces of water, and introduce them into the gauze-covered ports on the vessel. 6. Assemble the vessel on the fermentation system, and connect the facilities for temperature, dO2, pH, and foam control as described in Subheading 3.3., steps 4–7. 7. Set the impeller speed to 250 rpm to aid temperature equilibration. Aseptically, introduce 20 mL deuterated 13C-glucose solution, 45 mL deuterated 15N-ammonium sulfate solution, 2.5 mL deuterated biotin solution, and 2.5 mL deuterated PTM1 via the inoculation port (see Note 19). 8. Calibrate the dO2 probe as described in Subheading 3.3., step 9.

28

Pickford and O’Leary

3.7. Uniform 2H,13C,15N-Labeling in a Fermenter 1. Activate the aeration feedback control as described in Subheading 3.4., step 1. 2. Inoculate with the 50-mL seed culture (see Subheading 3.5., step 3) and start the controller to monitor dO2, impeller speed, temperature, and pH. Take a 10-mL sample via the sample port for assay and determination of wet cell weight and/or OD600 (see Note 11). 3. Aseptically, fill the methanol feed aspirator with the 2H,13C-methanol feed solution and connect the aspirator to the vessel via the nutrient-feed pump. 4. When the dO2 “spikes” (typically 36–48 h after inoculation), take a 10-mL sample for analysis, and add 50 mL deuterated 2H,13C-methanol feed solution (toxic) via the inoculation port (see Note 20). Start the methanol feed at 1.25 mL/h, increasing to 2.5 mL/h after 6 h. Take samples regularly (every 16–24 h) for analysis of cell density (OD600) and of expression level by SDS-PAGE (see Note 11). Continue the fermentation until it has consumed at least 250 mL methanol feed solution (see Note 3). 5. See Subheading 3.4., step 5. 6. Harvest the expressed protein by centrifugation as described in Subheading 3.2., step 5, and start purification immediately to minimize proteolytic degradation (see Subheading 3.8. and Note 14).

3.8. Protein Purification by Cation-Exchange Chromatography The following protocol is a convenient first-step procedure that serves not only to partially purify the recombinant protein but also to concentrate the sample volume for subsequent steps. However, if the heterologous protein is acid labile, insoluble at low pH, or is highly acidic (has a low isoelectric point) an alternative purification method should be used. 1. Dilute the crude supernatant twofold with Milli-Q water. Adjust the pH to 3.0 with 2 M hydrochloric acid (corrosive) or 2 M sodium hydroxide (corrosive). 2. Equilibrate the cation-exchange column of SP-Sepharose FF with 5 column volumes of cation-exchange equilibration buffer. 3. Apply the diluted protein sample onto the column at 10–20 mL/min. Collect the yellow-green nonbinding material that contains alcohol oxidase (AOX) and other acidic proteins (see Note 12). 4. Wash the column with 2 column volumes of cation-exchange equilibration buffer. 5. Elute the bound proteins by applying a gradient of 0–100% cation-exchange elution buffer and collecting fractions equivalent to 0.1 column volumes. 6. Analyze the fractions and the nonbinding material by SDS-PAGE, and pool those containing the recombinant protein.

3.9. Glycoprotein Purification by Lectin Affinity Chromatography Separation of glycosylated from nonglycosylated proteins secreted from P. pastoris can be achieved using affinity chromatography with immobilized jack

Isotopic Labeling of Pichia pastoris Proteins

29

bean Con A lectin. Con A has specificity for branched mannoses and carbohydrates with terminal mannose or glucose, and therefore, it binds to both N- and O-linked oligosaccharides produced by P. pastoris. 1. Regenerate a column of Con A-Sepharose (to remove any bound glycoproteins or sugar analogs) by washing with 10 column volumes of Con A regeneration buffer, followed by 10 column volumes of Con A equilibration buffer. 2. Add CaCl2, MgCl2, and MnCl2 to the protein sample, each to a final concentration of 1 mM. Adjust the sample pH to 7.4 with solid Tris base (see Note 22). 3. Apply the protein sample to the column at 0.1–0.5 mL/min, and wash the column with 5 column volumes of Con A equilibration buffer. Collect the nonbinding fraction that contains the nonglycosylated proteins. 4. To elute the glycoproteins and free oligosaccharides, wash the column with 5 column volumes of Con A elution buffer. Analyze the eluted glycoproteins and the nonbinding aglycosyl proteins by SDS-PAGE.

3.10. Enzymatic Deglycosylation of Glycoproteins The large hydrodynamic radius of N-linked, high-mannose sugars on recombinant proteins greatly increases their rotational correlation time (τc) leading to extensive line-broadening in their NMR spectra. In the majority of cases, the sugar moiety is not essential for structural integrity or functional activity of the protein and can be trimmed using the enzyme endoglycosidase (Endo) Hf (see Note 23). Endo Hf cleaves the chitobiose core of high-mannose and hybrid type N-linked oligosaccharides to leave a single N-linked N-acetylglucosamine (GlcNAc). Although enzymatic removal of O-linked oligosaccharides is possible and has been described in detail elsewhere (16), in our experience it is not economically viable to do so on the quantity of protein required for an NMR sample. 1. Adjust the protein sample to pH 5.5 with 1 M sodium hydroxide (corrosive). 2. Add 1 kU of Endo Hf (supplied by New England Biolabs) per milliliter of glycoprotein solution. Incubate at 37ºC for 1–3 h. 3. Analyze the glycosylated and aglycosyl protein by SDS-PAGE. N-linked glycosylation results in an increase in apparent molecular weight of approx 7 kDa per high-mannose chain, and a significant smearing of the protein band. 4. The removed oligosaccharides may be separated from the deglycosylated protein by Con A affinity chromatography (see Subheading 3.9.).

4. Notes 1. Selection of the growth and induction pH of P. pastoris can help to overcome proteolysis and insolubility of the recombinant product. The optimum value should be determined in small-scale cultures prior to isotopic labeling. P. pastoris is capable of growing across a relatively broad pH range from 3.0 to 7.0, which allows considerable leeway in adjusting the pH to one that is not optimal for a problem protease.

30

Pickford and O’Leary

2. Solutions of ammonium sulfate and D-glucose can be sterilized by autoclaving. Filter-sterilization is used instead as a precaution against accidental loss of the isotopically labeled compounds. D2O solutions should never be sterilized by autoclaving. 3. The amounts of 15N-,13C-, and 2H,13C-labeled isotopes given here have resulted in good yields in our experience, but given their high cost, it is highly recommended that a “dry run” be performed using unlabeled ammonium sulfate, D-glucose, and methanol to establish the minimum quantities needed for the required expression level. 4. Final adjustment of the fermentation media pH is performed after autoclaving as inorganic salts may precipitate during the sterilization process if the BSM pH is >3.5. Brady et al. (17) have solved this problem by reducing the concentration of all salts in the BSM to one-quarter of those given here without any adverse effect on cell growth rate, biomass yield, or the level of their recombinant protein. The reduced ionic strength may be beneficial when using ion-exchange chromatography as a first step in protein purification. 5. Different molecular-weight preparations of polypropylene glycol are available; we routinely use 1025 obtained from Merck (Poole, UK). The disadvantage of polypropylene glycol as an antifoam reagent is that its presence, even at low concentrations, can significantly reduce flow rates through certain ultrafiltration membranes. The addition of antifoam is usually necessary, but when growing P. pastoris at relatively low cell densities only small quantities are required. If the methanol feed is continued for longer periods, more antifoam solution may be needed. If the culture foams excessively, it is a sign of carbon source limitation, low pH, or ill health of the culture. If automated antifoam addition is unavailable, add the antifoam manually, but use the minimum amount to control foaming. 6. In general, no acid is required for Pichia fermentation because a healthy culture always acidifies the media. If the pH of the culture is increasing, it is a sign of carbon source depletion or ill health of the culture. 7. An OD600 = 1 is equivalent to 5 × 107 P. pastoris cells. 8. The volume of seed culture can be scaled according to the volume of the fermentation. In our experience, good results are obtained from using a seed culture that is 10% of the fermentation volume. 9. Aeration, particularly during the induction phase, is the most important parameter for efficient expression in shake flasks. Baffled flasks should always be used, and the volume of the culture should not exceed 30% of the flask volume. 10. Typical induction times range from 36 to 72 h, but will depend on the desired expression level, the stability of the expressed product, and the amount of 13C- or 2H,13C-methanol feed solution available. Optimal induction times should be determined from SDS-PAGE analysis of culture samples taken during the extended induction phase of a “dry run” using unlabeled D-glucose and methanol. 11. Regular sampling is important to monitor the growth of the fermentation, allows assay for any recombinant product, and permits microscopic inspection of the culture for contaminating organisms. Any precipitated inorganic salts in the sample

Isotopic Labeling of Pichia pastoris Proteins

12.

13.

14. 15. 16.

17. 18.

19. 20.

21.

31

can interfere with these tests but can be dissolved by the addition of an equal volume of 100 mM HCl without leading to cell lysis. If the recombinant product is acid labile, then a sample of supernatant should be removed for assay before addition of the HCl. The yellow-green color of the culture supernatant is the result of AOX that, although not actively secreted, accumulates in the media because of leakage from the cells. The AOX can be conveniently purified away from the recombinant material by cation-exchange chromatography (see Subheading 3.8.). The purpose of washing the cells is threefold: (a) to increase the yield of recombinant protein; (b) to dilute the supernatant and, therefore, to reduce its ionic strength prior to ion-exchange chromatography; and (c) to reduce the pH to approx 3.0 for cation exchange. If the recombinant protein is acid sensitive, then the cells should be washed with sterile H2O instead of 100 mM HCl. To minimize proteolysis in the crude supernatant, add phenylmethylsulfonyl fluoride (PMSF) (toxic) and ethylenediaminetetraacetic acid to a concentration of 1 mM. The PMSF should be refreshed every few hours. The performance of the gel-filled autoclavable pH probes will deteriorate with repeated sterilization but can be prolonged by periodic cleaning as recommended by the manufacturer. Probes should be stored in 3 M KCl. We have only used polarographic oxygen probes; information on the use of galvanic oxygen probes should be obtained from the manufacturer. Polarographic oxygen probes require routine maintenance to prolong their working life, and these details should also be obtained from the manufacturer. It takes approx 6 h to completely polarize a dO2 electrode. The polarization module applies the same voltage to the dO2 electrode as the controller. Autoclavable modules offer the speed advantage of performing electrode polarization during sterilization. The fermenter impeller speed can be controlled manually, but this is not recommended given the extended duration of P. pastoris fermentations. During the induction phase, the dO2 level must remain above 15% at all times to prevent the buildup of methanol in the vessel, but it must also stay below 50% to avoid methionine oxidation of the recombinant protein. As the pH rises above 3.5, and particularly on addition of PTM1 salts, some inorganic salts will precipitate. This does not affect cell growth. When a culture is growing normally and consumes the entire carbon source, respiration will slow dramatically, causing the dO2 level to rise sharply or “spike.” Addition of more carbon source should lead to an almost instantaneous rise in respiration signaled by a concomitant fall in dO2. Many fermenter controllers (including ours) have automated feedback control, which adjusts the impeller speed to maintain the dO2 at a preset level. This is very convenient when the system is left unattended for long periods of time but can mask dO2 fluctuations. In this case, test for carbon limitation by fixing the impeller speed at the current variable level and watch for dO2 “spikes.” The dO2 “spike” method of ensuring growth-limiting methanol concentration is labor intensive and repeatedly exposes the cells to noninducing levels of methanol.

32

Pickford and O’Leary

This can be avoided through the use of a methanol probe that allows continuous measurement and control of the methanol concentration in the media. 22. Con A is a metalloprotein that contains four metal binding sites. To maintain the binding characteristics of Con A, both Mn2+ and Ca2+ must be present in the binding buffer and protein sample. 23. Endo Hf is a protein fusion of Endo H (cloned from Streptomyces plicatus) and maltose binding protein. It has identical activity to Endo H but is less expensive. PNGase F may be used as an alternative to Endo Hf. This enzyme cleaves between the innermost GlcNAc and asparagine residues of high mannose, hybrid, and complex oligosaccharides from N-linked glycoproteins. However, the activity of this enzyme is more susceptible to steric hindrance of the scissile bond.

Acknowledgments The authors thank the Wellcome Trust for financial assistance and Prof. Iain D. Campbell and Dr. A. K. Downing for their support and direction. References 1. Higgins, D. R. and Cregg, J. M. (1998) Introduction to Pichia pastoris. In Methods in Molecular Biology, vol. 103: Pichia Protocols (Higgins, D. R. and Cregg, J. M., eds.), Humana, Totowa, NJ, pp. 1–15. 2. Cregg, J. M., Shen, S., Johnson, M., and Waterham, H. R. (1998) Classical genetic manipulation. In Methods in Molecular Biology, vol. 103: Pichia Protocols (Higgins, D. R. and Cregg, J. M., eds.), Humana, Totowa, NJ, pp. 17–26. 3. 3 Lin Cereghino, G. P., Lin Cereghino, J., Ilgen, C., and Cregg, J. M. (2002) Production of recombinant proteins in fermenter cultures of the yeast Pichia pastoris. Curr. Opin. Biotechnol. 13, 1–4. 4. 4 Grinna, L. S. and Tschopp, J. F. (1989) Size distribution and general structural features of N-linked oligosaccharides from the methylotrophic yeast, Pichia pastoris. Yeast 5, 107–115. 5. 5 Cregg, J. M., Vedvick, T. S., and Raschke, W. C. (1993) Recent advances in the expression of foreign genes in Pichia pastoris. Biotechnology 11, 905–910. 6. 6 Laroche, Y., Storme, V., De Meutter, J., Messens, J., and Lauwereys, M. (1994) High level secretion and very efficient isotopic labelling of tick anticoagulant peptide (TAP) expressed in the methylotrophic yeast Pichia pastoris. Biotechnology 12, 1119–1124. 7. 7 Cregg, J. M., Barringer, K. J., Hessler, A. Y., and Madden, K. R. (1985) Pichia pastoris as a host system for transformations. Mol. Cell. Biol. 5, 3376–3385. 8. 8 Scorer, C.A., Clare, J. J., McCombie, W. R., Romanos, M. A., and Sreekrishna, K. (1994) Rapid selection using G418 of high-copy number transformants of Pichia pastoris for high-level foreign gene expression. Biotechnology 12, 181–184. 9. 9 Romanos, M. A. (1995) Advances in the use of Pichia pastoris for high-level gene expression. Curr. Opin. Biotechnol. 6, 527–533.

Isotopic Labeling of Pichia pastoris Proteins

33

10. Invitrogen Corporation. Pichia Expression System: Manual of Methods for Expression of Recombinant Proteins in Pichia pastoris, Version G. Available at http://www.invitrogen.com/content/sfs/manuals/pich_man.pdf. Accessed 03/01/04. 11. Stratton, J., Chriuvolu, V., and Meagher, M. (1998) High cell-density fermentation. In Methods in Molecular Biology, vol. 103: Pichia Protocols (Higgins, D. R. and Cregg, J. M., eds.), Humana, Totowa, NJ, pp. 107–120. 12. Invitrogen Corporation. Pichia Fermentation Process Guidelines. Available at http://www.invitrogen.com/content/sfs/manuals/pichiaferm_prot.pdf. Accessed 03/01/04. 13. 13 Clare, J. J., Romanos, M. A., Rayment, F. B., Rowedder, J. E., Smith, M. A., Payne, M. M., et al. (1991) Production of mouse epidermal growth factor in yeast: highlevel secretion using Pichia pastoris strains containing multiple gene copies. Gene 105, 205–212. 14. Gleeson, M. A. G., White, C. E., Meininger, D. P., and Komives, E. A. (1998) Generation of protease-deficient strains and their use in heterologous protein expression. In Methods in Molecular Biology, vol. 103: Pichia Protocols (Higgins, D. R. and Cregg, J. M., eds.), Humana, Totowa, NJ, pp. 81–94. 15. Bright, J. R., Pickford, A. R., Potts, J. R., and Campbell, I. D. (1999) Preparation of isotopically labelled recombinant fragments of fibronectin for functional and structural study by heteronuclear nuclear magnetic resonance spectroscopy. In Methods in Molecular Biology, vol. 139: Extracellular Matrix Protocols (Streuli, C. and Grant, M., eds.), Humana, Totowa, NJ, pp. 59–69. 16. Cremata, J. A., Montesino, R., Quintero, O., and Garcia, R. (1998) Glycosylation profiling of heterologous proteins. In Methods in Molecular Biology, vol. 103: Pichia Protocols (Higgins, D. R. and Cregg, J. M., eds.), Humana, Totowa, NJ, pp. 95–105. 17. Brady, C. P., Shimp, R. L., Miles, A. P., Whitmore, M., and Stowers, A. W. (2001) High-level production and purification of P30P2MSP119, an important vaccine antigen for malaria, expressed in the methylotrophic yeast Pichia pastoris. Protein Expr. Purif. 23, 468–475.

3 Perdeuteration/Site-Specific Protonation Approaches for High-Molecular-Weight Proteins Stephen Matthews Summary Among the factors that limit the application of nuclear magnetic resonance (NMR) to biological macromolecules are increasing resonance overlap and fast transverse relaxation. Multidimensional NMR combined with 13C and 15N labeling has alleviated these problems temporarily; however, they resurface at molecular weight (mol wt) in excess of 30 kDa. Combined perdeuteration/site-specific protonation together with segmental labeling (see Chapter 4), transverse relaxation-optimized spectroscopy (TROSY) (see Chapter 5), and residual dipolar couplings (see Chapter 7) have all helped to dramatically extend the mol wt limit. This article describes some of the practical aspects of the combined perdeuteration/site-specific protonation approach, which has proved so useful in the global fold determination of large proteins. Key Words: Perdeuteration; site-specific protonation; large proteins; global folds; residual dipolar couplings.

1. Introduction The deuteration of molecules for the simplification of nuclear magnetic resonance (NMR) spectra is not a new strategy, but its combination with tripleresonance spectroscopy and site-specific protonation has generously extended the molecular size limitation for NMR applications to macromolecules. The deuteron has a significantly smaller magnetogyric ratio (1/6.5 that of a proton), and consequently deuteration removes the dominant 1H-13C dipolar relaxation mechanism for 13C nuclei. Typically, the transverse relaxation of the Cα is approximately fivefold slower in deuterated proteins. This, in combination with 2H-decoupling during 13C evolution, ensures that triple-resonance experiments, namely HNCA, HNCOCA, HNCACB, HNCOCACB, and HNCACO, work very efficiently on proteins in excess of 30 kDa (1,2). In addition, the cross-peak

From: Methods in Molecular Biology, vol. 278: Protein NMR Techniques Edited by: A. K. Downing © Humana Press Inc., Totowa, NJ

35

36

Matthews

phase in the double-constant time d-HN(CA)CB and d-HN(COCA)CB experiments is particularly useful for rapid sequential assignment (1). The 13Cβ nuclei that are coupled to an odd number of aliphatic carbons give one sign and 13C nuclei that are coupled to an even number of aliphatic carbons give the β opposite sign. Transverse relaxation in deuterated proteins above 60 kDa becomes sufficiently fast that the standard triple-resonance approaches to assignment begin to fail. Pervushin and coworkers have developed a new class of NMR experiment that offers a further relaxation benefit in deuterated proteins (3). Transverse relaxation-optimized spectroscopy (TROSY) selects transitions that experience mutual cancellation of dipole–dipole and chemical shift anisotropy relaxation mechanisms. Furthermore, any residual relaxation because of dipolar interactions with remote protons can almost be entirely removed by perdeuteration. The result is a dramatic decrease in resonance line-widths that enables NMR spectra to be obtained for proteins in excess of 100 kDa. Numerous improvements and derivatives have been published (4,5), with perhaps the current highlight being the application of cross-correlated relaxation-enhanced polarization transfer and cross-correlated relaxation-induced polarization transfer to 470 and 800 kDa GroEL/GroES complexes (6). These TROSY-based elements have also been incorporated into the repertoire of multidimensional assignment experiments and have further extended their applicability (7). Notably, Tugarinov et al. have applied these techniques successfully to the 723-residue protein, malate synthase G (8). One of the limiting factors of perdeuteration is loss of protons between which nuclear Overhauser effect (NOE) measurements can be made and the subsequent impact this has on the quality of resulting structures. Mal et al. demonstrated on Fn3 that the longer distances measurable in deuterated macromolecules facilitate the characterization of the fold for an exclusively β-sheet protein using reprotonated amides alone (9). α-helical proteins possess many fewer long-range amide–amide NOEs and, therefore, are not amenable to this approach. Gardner et al. established the limitations of such an approach using an SH2 domain from phospholipase C γ1 (10). To increase the number of NOEs observed, several groups have assessed the feasibility of specific protonation of amino acids within a deuterium background (11,12). Amino acids containing methyl groups were chosen for maximum impact because they provide improved line-widths, fall in well-resolved regions of the spectrum, and are common within protein hydrophobic cores. Rosen et al. described a novel route to achieving methyl protonated samples that used 13C,1H-pyruvate as the carbon source rather than supplementing D2O-based growth media with the protonated amino acids (13). By far the most elegant method was that of Gardner et al., which involved the selective reintroduction of protons into the methyl group

High-Molecular-Weight Proteins

37

positions of valine, leucine, and isoleucine (10,14–16). Escherichia coli will readily use the biosynthetic precursors of these amino acids, namely [3,3-2H2] 13C-2-ketobuterate and 13C-α-ketoisovalerate, in preference to complete biosynthesis. The final product is a deuterated protein that is specifically protonated at the terminal methyl groups of valine, leucine, and isoleucine. Simulations suggest that the numerous NOE correlations between methyl groups within the hydrophobic core are sufficient to confidently characterize the global fold of many, if not all, proteins to within a few angstrom root-meansquare deviation (RMSD) (10). In cases where higher precision is required, one strategy is to introduce additional protons, with perhaps aromatic sidechains being the most likely to provide substantial benefits (12). The strong 13C–13C J-coupling and poor chemical shift dispersion can greatly hinder the assignment of NOEs involving the aromatic rings, particularly phenylalanine. Wang and coworkers proposed a method of selectively labeling the ε-proton in phenylalanine sidechains, which, when combined with selective methyl group labeling, would improve the precision of structures considerably (17). The introduction of residual dipolar couplings (RDCs) can afford further improvements in the quality of structures derived from partially deuterated material. Moreover, RDCs are highly applicable to deuterated molecules as they can be measured between nonproton nuclei, that is, 15N-13C and 13C-13C. The practical aspects of RDC applications are covered in Chapter 7 of this book and are not addressed in this chapter, although they represent a natural extension to this work. 2. Materials 2.1. Deuterium and

13C/15N

Minimal Media

M9 salts: 6.0 g Na2HPO4, 3.0 g KH2PO4, 0.5 g NaCl, 0.7 g 15NH4Cl, 1 L D2O. 20% 13C/1H glucose. 1 M MgSO4. 0.1 M CaCl2. 1 M thiamine. E. coli trace elements, pH 7.0; 5 g ethylenediaminetetraacetic acid, 800 mL D2O, 0.5 g FeCl3, 0.05 g ZnCl2, 0.01 g CuCl2, 0.01 g CoCl2.6H2O, 0.01 g H3BO3. 7. 100 mg/mL antibiotic solution.

1. 2. 3. 4. 5. 6.

2.2. Specific Protonation 1. 50 mg [3,3-2H2] α-ketobutyrate (Cambridge Isotope Laboratory, Andover, MA, cat. no. CDLM4611). 2. 100 mg 15N, 13C-[2,3-2H2]-valine (Cambridge Isotope Laboratory, Andover, MA, cat. no. CDNLM4281) or 75 mg 13C α-ketoisovalerate. 3. 30 mg ε-13C-phenylalanine.

38

Matthews

2.3. NMR Spectroscopy and Structure Calculation 1. Four-channel pulsed-field gradient NMR spectrometer equipped with deuterium decoupling capabilities. 2. NMRview (18) or other suitable NMR assignment software. 3. XPLOR/CNS program (19–22).

3. Methods The methods described in this section outline preparation, NMR spectroscopy, and structure calculation of deuterated proteins with specific protonation. 3.1. Preparation of Deuterated Proteins With Specific Protonation DNA encoding Int280 (EPEC strain E2348/69) was cloned into the expression vector pET3d and freshly transformed into a strain of BL21 pLysS E. coli (23,24). 1. Prepare 1 L of D2O minimal culture medium (see Note 1). 2. Add approx 1/10 of labeled amino acid (i.e., 5 mg [3,3-2H2] α-ketobutyrate, 10 mg 15N, 13C-[2,3-2H ]-valine and/or 3 mg ε-13C-phenylalanine) (see Note 2). 2 3. Filter-sterilize culture medium. 4. Inoculate 1 L of D2O minimal culture media and grow at 37ºC. 5. At 1 h prior to induction add the rest of the labeled amino acid(s) (i.e., 45 mg [3,3-2H2] α-ketobutyrate, 90 mg 15N, 13C-[2,3-2H2]-valine and/or 27 mg ε-13Cphenylalanine) (see Note 2). 6. Induce at optical density (OD)600 approx 0.6 by adding isopropyl-β-D-thiogalactopyranoside to 120 mg/L, and grow for at least 5 h before harvesting. 7. Disrupt cells, purify protein, and prepare for NMR spectroscopy.

3.2. NMR Spectroscopy Pure protein was extensively dialyzed against 20 mM sodium acetate buffer at pH 5.2, concentrated to 1 mM in 300 µL, and placed in a 5-mm Shigemi tube (Shigemi, Allison Park, PA). All NMR spectra were recorded at 500 MHz proton frequency on a four-channel Bruker DRX500 equipped with a z-shielded gradient, triple-resonance probe. NMR data were processed using XWINNMR and NMRPipe (25) and analyzed using NMRview (18). The temperature was maintained at 310 K throughout the experiments (23,24). One-dimensional and two-dimensional (2D) NMR spectra of site-specifically protonated, deuterated Int280 are shown in Figs. 1 and 2. 1. Determine sequence-specific backbone 1HN, 15N, 13Cα, 13Cβ, and C′ assignments using the standard triple-resonance approach (1,2) (see Note 3). 2. Assign valine, leucine, and isoleucine methyl group resonances using (H)C(CO)NH-total correlation spectroscopy (TOCSY) (1) (see Note 4).

High-Molecular-Weight Proteins

39

Fig. 1. 1D 1H NMR spectra of (A) 1H,13C,15N Int280, (B) 2H,13C,15N Int280 showing the dramatic improvement in amide proton line-widths on deuteration, and (C) 13CH –Val, Leu,2H,13C,15N Int280, illustrating the selectivity of methyl group labeling. 3

3. Assign NOEs for use in structure calculation from 1H/15N- and 1H/13C-edited NOE heteronuclear single-quantum coherence (NOESY–HSQC) spectroscopy experiments (26,27), as well as 3D/4D 15N/15N-, 15N/13C- and 13C/13C-edited HMQC–NOESY–HSQC experiments (28) (see Note 5). Two mixing times of 100 and 200 ms were chosen for this protein, which has a correlation time of 16 ns. 4. Identify amide protons involved in hydrogen bonds by their resistance to exchange in the presence of D2O. NH resonances observed in HSQC spectra recorded after dissolving or buffer exchanging into 100% D2O are assigned as hydrogen bond donors.

3.3. Structure Calculation 1. Measure NOE cross-peak intensities at both 100 and 200 ms mixing times (see Note 6). 2. Calibrate NOE-based distance restraints using known sequential distances, that is, NH–NH(i,i+1) in the center of a helix.

40

Matthews

Fig. 2. Methyl group region of the 2D 1H-13C CT-HSQC NMR spectrum of Leu,2H,13C,15N Int280 illustrating the high quality of data and selectivity of methyl group labeling. 13CH –Val, 3

3. Place NOEs observed at a mixing time of 100 ms into two categories on the basis of estimated NOE cross-peak intensity: strong ( 4) connectivities, of which a total of 502 involved methyl groups. One hundred and ten dihedral angle restraints were used for initial calculations of the fold, but were subsequently removed during refinement. The orientation of D1 with respect to the rest of the structure is ill defined. To illustrate precision of the D1 family, it has been superimposed separately on a representative structure. RMSDs are 1.5, 1.8, and 1.8 Å for backbone atoms position in D1, D2 and D3, respectively.

the 370-residue maltodextrin-binding protein (10), 300 ms for the α-helical, antiapoptotic protein Bcl-xL (32), 175 ms for Dbl homology domain (33), and 200 ms deuterated OmpA in micelles (34). 7. The 13Cα, 13Cβ and 13C' shifts were corrected for deuterium isotope effects using the method described by Venter et al. (35). In cases where angles found by TALOS do not satisfy the acceptance criteria, chemical shift index predictions can be used (36).

High-Molecular-Weight Proteins

43

Acknowledgments S. Matthews would like to acknowledge the support of the Wellcome Trust and the BBSRC. The atomic coordinates for Int280 have been deposited in the Brookhaven PDB with accession number 1INM and assignments in BioMag–ResBank under the accession number 4111. References 1. 1 Shan, X., Gardner, K. H., Muhandiram, D. R., Rao, N. S., Arrowsmith, C. H., and

2.

3. 3

4. 4 5. 5 6. 6 7. 7

8. 8 9. 9 10. 10 11. 11

Kay, L. E. (1996) Assignment of 15N, 13Cα, 13Cβ and HN resonances in an 15N, 13C, 2H labeled 64 kDa trp repressor–operator complex using triple resonance NMR spectroscopy and 2H-decoupling. J. Am. Chem. Soc. 118, 6570–6579. Yamazaki, T., Lee, W., Arrowsmith, C. H., Muhandiram, D. R., and Kay, L. (1994) A suite of triple resonance NMR experiments for the backbone assignment of 15N, 13C, 2H labeled proteins with high sensitivity. J. Am. Chem. Soc. 116, 11,655–11,666. Pervushin, K., Riek, R., Wider, G., and Wuthrich, K. (1997) Attenuated T-2 relaxation by mutual cancellation of dipole–dipole coupling and chemical shift anisotropy indicates an avenue to NMR structures of very large biological macromolecules in solution. Proc. Natl. Acad. Sci. USA 94, 12,366–12,371. Riek, R., Pervushin, K., and Wuthrich, K. (2000) TROSY and CRINEPT: NMR with large molecular and supramolecular structures in solution. Trends Biochem. Sci. 25, 462–468. Riek, R., Fiaux, J., Bertelsen, E. B., Horwich, A. L., and Wuthrich, K. (2002) Solution NMR techniques for large molecular and supramolecular structures. J. Am. Chem. Soc. 124, 12,144–12,153. Fiaux, J., Bertelsen, E. B., Horwich, A. L., and Wuthrich, K. (2002) NMR analysis of a 900K GroEL–GroES complex. Nature 418, 207–211. Salzmann, M., Pervushin, K., Wider, G., Senn, H., and Wuthrich, K. (2000) NMR assignment and secondary structure determination of an octameric 110 kDa protein using TROSY in triple resonance experiments. J. Am. Chem. Soc. 122, 7543–7548. Tugarinov, V., Muhandiram, R., Ayed, A., and Kay, L. E. (2002) Four-dimensional NMR spectroscopy of a 723-residue protein: chemical shift assignments and secondary structure of malate synthase G. J. Am. Chem. Soc. 124, 10,025–10,035. Mal, T. K., Matthews, S. J., Kovacs, H., Campbell, I. D., and Boyd, J. (1998) Some NMR experiments and a structure determination employing a {N-15,H-2} enriched protein. J. Biomol. NMR 12, 259–276. Gardner, K. H., Rosen, M. K., and Kay, L. E. (1997) Global folds of highly deuterated, methyl-protonated proteins by multidimensional NMR. Biochemistry 36, 1389–1401. Metzler, W. J., Wittekind, M., Goldfarb, V., Mueller, L., and Farmer, B. T. (1996) Incorporation of 1H/13C/15N-(Ile, Leu, Val) into a perdeuterated, 15N-labeled protein: potential in structure determination of large proteins by NMR. J. Am. Chem. Soc. 118, 6800, 6801.

44

Matthews

12. Smith, B. O., Ito, Y., Raine, A., Teichmann, S., Ben-Tovim, L., Nietlispach, D., et al. (1996) An approach to global fold determination using limited NMR data from larger proteins selectively protonated at specific residues. J. Biomol. NMR 8, 360–368. 13. 13 Rosen, M. K., Gardner, K. H., Willis, R. C., Parris, W. E., Pawson, T., and Kay, L. E. (1996) Selective methyl group protonation of perdeuterated proteins. J. Mol. Biol. 263, 627–636. 14. 14 Gardner, K. H. and Kay, L. E. (1997) Production and incorporation of N-15, C-13, H-2 (H-1-delta 1 methyl) isoleucine into proteins for multidimensional NMR studies. J. Am. Chem. Soc. 119, 7599, 7600. 15. 15 Zwahlen, C., Vincent, S. J. F., Gardner, K. H., and Kay, L. E. (1998) Significantly improved resolution for NOE correlations from valine and isoleucine (C-gamma 2) methyl groups in N-15,C-13- and N-15,C-13,H-2-labeled proteins. J. Am. Chem. Soc. 120, 4825–4831. 16. 16 Goto, N. K., Gardner, K. H., Mueller, G. A., Willis, R. C., and Kay, L. E. (1999) A robust and cost-effective method for the production of Val, Leu, Ile (delta 1) methyl-protonated N-15-, C-13-, H-2-labeled proteins. J. Biomol. NMR 13, 369–374. 17. 17 Wang, H., Janowick, D. A., Schkeryantz, J. M., Liu, X., and Fesik, S. W. (1999) A method for assigning phenylalanines in proteins. J. Am. Chem. Soc. 121, 1611, 1612. 18. Johnson, B. A. and Blevins, R. A. (1994) NMRView: a computer program for the visualization and analysis of NMR data. J. Biomol. NMR 4, 603–614. 19. 19 Stein, E., G., Rice, L. M., and Brünger, A. T. (1997) Torsion-angle molecular dynamics as a new efficient tool for NMR structure calculation. J. Magn. Reson. B 124, 154–164. 20. 20 Nilges, M., Gronenborn, A. M., and Clore, G. M. (1988) Determination of 3dimensional structures of proteins by simulated annealing with interproton distance restraints—application to crambin, potato carboxypeptidase inhibitor and barley serine proteinase inhibitor-2. Protein Eng. 2, 27–38. 21. Brünger, A. T. (1992) XPLOR Manual Ver. 3.1. Yale University, New Haven, CT. 22. Brünger, A. T., Adams, P. D., Clore, G. M., DeLanod, W. L., Grose, P., GrosseKunstleve, R. W., et al. (1998) Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta. Crystallogr. D Biol. Crystallogr. D54, 905–921. 23. 23 Kelly, G., Prasannan, S., Daniell, S., Frankel, G., Dougan, G., Connerton, I., et al. (1998) Sequential assignment of the triple labeled 30.1 kDa cell-adhesion domain of intimin from enteropathogenic E-coli. J. Biomol. NMR 12, 189–191. 24. Kelly, G., Prasannan, S., Daniell, S., Fleming, K., Frankel, G., Dougan, G., et al. 24 (1999) Structure of the cell-adhesion fragment of intimin from enteropathogenic Escherichia coli. Nat. Struct. Biol. 6, 313–318. 25. 25 Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A. (1995) NMRPipe: a multidimensional spectral processing system based on Unix pipes. J. Biomol. NMR 6, 277–293.

High-Molecular-Weight Proteins

45

26. Norwood, T. J., Boyd, J., Heritage, J. E., Soffe, N., and Campbell, I. D. (1990) Comparison of techniques for 1H-detected heteronuclear 1H-15N spectroscopy. J. Magn. Reson. B 87, 488–501. 27. 27 Marion, D., Driscoll, P. C., Kay, L. E., Wingfield, P. T., Bax, A., Gronenborn, A., et al. (1989) Overcoming the overlap problem in the assignment of larger proteins by the use of three-dimensional heteronuclear 1H-15N Hartmann-Hahn-multiple quantum coherence and nuclear Overhauser multiple quantum coherence spectroscopy: application to interleukin 1. Biochemistry 28, 6150–6156. 28. 28 Vuister, G. W., Clore, G. M., Gronenborn, A. M., Powers, R., Garrett, D. S., Tschudin, R., et al. (1993) Increased resolution and improved spectral quality in 4dimensional 13C/13C-separated HMQC-NOESY-HMQC spectra using pulsed field gradients. J. Magn. Reson. B 101, 210–213. 29. 29 Cornilescu, G., Delaglio, F., and Bax, A. (1999) Protein backbone angle restraints from searching a database for chemical shift and sequence homology. J. Biomol. NMR 13, 289–302. 30. 30 Goto, N. K. and Kay, L. E. (2000) New developments in isotope labeling strategies for protein solution NMR spectroscopy. Curr. Opin. Struct. Biol. 10, 585–592. 31. 31 Mulder, F. A. A., Ayed, A., Yang, D. W., Arrowsmith, C. H., and Kay, L. E. (2000) Assignment of H-1(N), N-15, C-13(alpha), (CO)-C-13 and C- 13(beta) resonances in a 67 kDa p53 dimer using 4D-TROSY NMR spectroscopy. J. Biomol. NMR 18, 173–176. 32. 32 Medek, A., Olejniczak, E. T., Meadows, R. P., and Fesik, S. W. (2000) An approach for high-throughput structure determination of proteins by NMR spectroscopy. J. Biomol. NMR 18, 229–238. 33. 33 Aghazadeh, B., Zhu, K., Kubiseski, T. J., Liu, G. A., Pawson, T. P., Zheng, Y., et al. (1998) Structure and mutagenesis of the Dbl homology domain. Nat. Struct. Biol. 5, 1098–1107. 34. 34 Arora, A., Abildgaard, F., Bushweller, J. H., and Tamm, L. K. (2001) Structure of outer membrane protein A transmembrane domain by NMR spectroscopy. Nat. Struct. Biol. 8, 334–338. 35. 35 Venter, R. A., Farmer, B. T., Fierke, C. A., and Spicer, L. D. (1996) Characterizing the use of perdeuteration in NMR studies of large proteins: C-13, N-15 and H-1 assignments of human carbonic anhydrase II. J. Mol. Biol. 264, 1101–1116. 36. Wishart, D. S. and Sykes, B. D. (1994) The 13C chemical shift index: a simple method for the identification of protein secondary structure using 13C chemical shift data. J. Biomol. NMR 4, 171–180.

4 Segmental Isotopic Labeling for Structural Biological Applications of NMR David Cowburn, Alexander Shekhtman, Rong Xu, Jennifer J. Ottesen, and Tom W. Muir Summary This chapter describes the preparation of precursor domains for the formation of multidomain segmentally labeled proteins by protein ligation. Key Words: Expressed protein ligation; inteins; isotopic labeling; reduced proton density labeling (REDPRO).

1. Introduction It is widely recognized that the composition of the human proteome has resulted from extensive domain duplication, with a significant proportion of the most duplicated of domain superfamilies being specific to multicellular function (1). Many of these domains within a particular protein may be independently folded and have a specific recognition function. The composite function of such a protein is nonetheless dependent on the organization of these domains and is rarely simply the sum of the parts. Although X-ray crystallography remains the gold standard for atomic resolution structures, the orientation of domains in crystal forms may be influenced by the crystal packing, and the interconversion between different ligated forms may be hard to study in the solid state. Thus the individual domain–domain interfaces require additional investigation by other methods. Nuclear magnetic resonance (NMR) can significantly play this role, especially with recent advances in improved sensitivity and extension of molecular weight limits (2). Even with the powerful transverse relaxation-optimized spectroscopy methods (3), there still remains the “too many notes” problem of multiple signals (4) and excess signal complexity, especially in nonsymmetric

From: Methods in Molecular Biology, vol. 278: Protein NMR Techniques Edited by: A. K. Downing © Humana Press Inc., Totowa, NJ

47

48

Cowburn et al.

Fig. 1. Thioester/N-terminal cysteine chemical ligation.

higher molecular weight proteins. Isotopic labeling of a segment of a protein permits a wide range of structural and mapping studies to be conducted by NMR (for a review, see ref. 5). The total synthesis of peptides, predominantly by solid state methods, is well developed and provides essentially complete control of the placement of isotopic labels, regio- and stereospecifically, but the economies and time-scales of such syntheses currently preclude their use for routine NMR analysis of products larger than about 3 kDa. The stitching together of segments of a protein labeled separately in bacterial or other expression systems provides a more general and effective route to segmental labeling. How is such stitching together to be done? The junction needs to resemble the natural peptidic linkage, and this

Segmental Isotopic Labeling

49

presents a major challenge to any synthetic method compatible with native expressed proteins. With synthetic products, the most effective chemical ligation method generally is accepted to be that of the reaction of a C-terminal thioester with an N-terminal cysteine residue (6). An overview of this reaction is in Fig. 1. This chemical reaction was subsequently discovered to be very similar to that used by nature in intein fusion of proteins known as protein splicing (7,8). The fascinating details of protein splicing are beyond the scope of this chapter and may be reviewed elsewhere (9,10). Subsequent molecular biology developments identified methods for the production of C-terminal α-thioesters and for the production of split inteins. These provide the base for the two current methods of segmental labeling—expressed protein ligation (EPL) and trans-splicing. The two methods are contrasted in Fig. 2. This chapter describes the use of EPL. Considerable success with the alternate trans-splicing method has been described (11–13). Our preference for the use of EPL is based on its ability to couple together native expressed fragments, without refolding, and its much less restrictive requirements on the linker composition. (see ref. 14). The major procedures involved in these protocols are continuously evolving— the general steps are illustrated in Fig. 3.

Fig. 2. Comparison of segmental labeling by EPL, and by trans-splicing.

50

Cowburn et al.

2. Materials Because the exact reagents required depend on some of the strategies for the above steps, we list only the key material requirements, especially for the reaction combination. There are multiple strategies for producing the two segments described in Subheading 1. Many projects will involve joining domains that have already been expressed, and these are likely to provide useful starting points with appropriate cloning sites and substituted codons for efficient expression. 1. C-Terminal Intein Fusion: Multiple vectors are commercially available from

New England Biolabs (Beverly, MA; www.neb.com). Many are described in Table 1 of ref. 14. The pTXB1 vector has been generally useful (e.g., ref. 15) in BL21(DE3) or derivatives of Escherichia coli. 2. N-Terminal Cys-Containing Protein: Commercial primers for N-terminal mutation/insertion are available from a large number of businesses and academic resource facilities. Examples include Integrated DNA Technologies (Coralville, IA; www.idtdna.com) and Invitrogen (Carlsbad, CA; www.invitrogen.com). For cells, it is our experience that frequent replenishment from a commercial source is desirable (e.g., Novagen, Madison, WI; www.novagen.com). Factor Xa for cleavage of precursor proteins available from Amersham Biosciences, Piscataway, NJ. 3. Equipment: The experimental approach outlined will require resources generally available in an NMR lab currently using genetic engineering and protein expression. Without being totally exhaustive, this list includes

Segmental Isotopic Labeling

51

analytical high-pressure (performance) liquid chromatography, preparative liquid chromatography, polymerase chain reaction (PCR) equipment, cellgrowth equipment (incubators, etc.), preparative centrifuge, concentrators, items required for the target’s functional assay, and gel electrophoresis. Required analytical services, possibly in a central resource, include protein mass spectrometry and DNA sequencing. This chapter does not deal with the NMR applications to the targets, described in other chapters, and the spectrometer equipment so required. 3. Methods 3.1. Overall Strategic Issues One of the major advantages of using EPL is that it can avoid refolding steps, but to do so does require that (a) one or both precursor segments are folded (as in ref. 16), (b) one of the precursor segments is folded and the other folds spontaneously when ligated (as in ref. 15), or (c) the precursor segments are not fully folded, but ligation induces their spontaneous fold. Conditions (a) and (b) are most readily identified with systems where significant “domainology” is already available, as has been the case in the work to date (15,16). 3.1.1. Position of the Ligation Point Comparison of domain behaviors should permit identification of likely linker regions between folded subdomains. It is desirable to maintain the same length of a linker and to substitute two or three residues in the final fusion at the junction site. When a linker is believed to be entirely flexible, it may be possible to insert additional residues at the junction site. Preexisting structural data, or homology modeling is likely to identify suitable areas, especially those containing hydrophilic and small side chain residues, which will be changed to −Gly-Cys-Gly- or −Gly-Cys- in the final fusion product. In other words, the “Domain-I” will have a C-terminal Gly preceding the α-thioester, and the “Domain-II” will have an N-terminal Cys, possibly followed by a Gly. The purpose of the first -Gly-, preceding the α-thioester, is to minimize steric hindrance of the transesterification of Fig. 2. The second Gly in the “Domain-II” facilitates the preparation of the N-terminal Cys Domain-II by cleavage using Factor Xa, and possibly other proteases. It is not obligatory even in these cases. Of course, one danger of these short peptide substitutions is that they may result in structural and/or functional modification related to the question under study. A conservative approach is then to prepare the mutant whole protein by conventional methods to test its relevant properties in advance of the vector preparation of the precursor sequences and isotopic labeling (see Note 1).

52

Cowburn et al.

3.1.2. Composition of the Ligation Surrounds The rate of ligation in Fig. 1 will be partly limited by the conformational flexibility of ligating ends of the sequences. To reduce this, it is wise to avoid inserting the linker point in any region likely to have secondary structure. In addition, very large clustering of similarly charged residues or of hydrophobic amino acids may lead to restricted conformational flexibility and reduce the rate of ligation. 3.2. Preparation of α-Thioester Protein 3.2.1. Expression and Purification of Protein–Intein Fusion Precursor For the preparation of the α-thioester protein, the vendor’s recommended cloning procedures and expression methods with chitin-binding domain (CBD)-mutated inteins are generally appropriate, as is the affinity chromatography of the expressed fusion protein on chitin-agarose beads. As a starting estimate, about 300 µg protein are likely to be absorbed per milliliter of chitin beads, and the volume of beads can be scaled to the estimated protein yield. The lysate from cell cleavage may be resuspended in a suitable column buffer (e.g., 20 mM sodium HEPES, 350 mM sodium chloride, 1 mM ethylenediaminetetraacetic acid, 0.1% Triton X-100, pH 7.0), and applied to the chitin bead column. The column is then washed with the same buffer and stored at 4ºC. Loading of this column may be determined by treating 100 µL of beads for 12 h with 200 mM sodium phosphate, 200 mM sodium chloride, 100 mM dithiothreitol, pH 7.2, washing the beads with 1:1 acetonitrile:water, and loading the solution and washes onto a high-pressure liquid chromatograph (HPLC) for quantitation. 3.2.2. α-Thioester Protein Production Fusion can be achieved either with or without intermediate isolation of the αthioester protein. In either case, it is probably worthwhile to ensure that coupling with a test peptide can be achieved with released α-thioester protein. A model peptide such as NH2-CGRGRGRK[fluorescein]-CONH2 (16) can be used to readily ensure that the construct designed can be released and coupled, prior to reaction with the domain-II product. The production of the fluorescein conjugated domain-I can then be monitored by the fluorescent band on gel electrophoresis. 3.2.3. Release of Thioester Protein Using Ethanethiol (Caution: See Note 2) Domain-I-intein-CBD fusion bound to chitin beads is equilibrated and suspended in 0.2 M sodium phosphate, pH 6.0, 0.2 M NaCl, and 3% (v/v) ethanethiol. The suspension is gently agitated overnight, the supernatant removed, and the beads washed with a releasing buffer, such as 1:1 (v/v) acetonitrile (HPLC grade)/water. All washes and the supernatant are combined and

Segmental Isotopic Labeling

53

loaded onto a retaining chromatographic system, such as C18 preparative HPLC, and purified. α-thioesters are highly reactive with aldehydes, and ketones—all glassware or plasticware should be prerinsed in buffers, and under no circumstances should glassware be dried with acetone (see Note 3). 3.2.4. Alternative Release Using 2-Mercaptoethanesulfonic Acid (MENSA) Production of stable α-thoiesters has also been reported using MENSA (17). 3.3. N-Terminal Cysteine-Containing Protein Several methods have been developed that include mutated inteins releasing an N-terminal cysteine protein (14). The locally preferred method uses Factor Xa to release the final product from a specific fusion system (15,16,18). The Nterminal insertion sequence −IEGRC- may be used. This contains the previously well-recognized −IEGR- motif for Factor Xa, with a C-terminal glycyl residue improving product yield (15). Alternative methods are available (19,20). 3.4. Final Product Preparation 3.4.1. Optimizing the Ligation Reaction The general limitation of the synthesis in Fig. 1 is that the speed of the second order reaction depends on the concentration of both precursors. This concentration may be limited by solubility or by the amount of material on hand. A large excess of one of the sequence reactants is a possible approach to this issue, but it will commonly be limited by lack of availability, especially for stable isotope-labeled materials. Some effort should be devoted to surveying a range of conditions of concentration, temperature, pH, and so forth, using analytical polyacrylamide electrophoresis of the products to assess the degree of product formed. As starting conditions for assessing ligation, the limiting component sequence should be in excess of 200 µM with buffer conditions of 200 mM sodium phosphate, pH 7.2, 200 mM sodium chloride, thiophenol (1.5% w/v), and benzyl mercaptan (1.5% w/v). 3.4.2. Isotopic Labeling Issues The general issues of isotopic labeling for NMR are well known (e.g., ref. 21). The most obvious use of segmental labeling is to characterize only one of the subsequent domains using labeling and leaving the other “cryptic.” More elaborate schemes are likely to be developed. 3.4.2.1. THE REDUCED PROTON (REDPRO) APPROACH

The REDPRO procedure has the potential for applications to structure determination, for mapping of chemical shift changes, and for application to higher

54

Cowburn et al.

molecular weight cases (22). Detailed descriptions of related procedures may also be consulted (23,24). It was previously observed (23) that E. coli BL21 can be grown without prior adaptation on minimal medium containing 100% 2D2O supplemented with [U-13C,1H] glucose and [U-15N,1H] ammonium chloride. E. coli BL21 (DE3) is freshly transformed before each expression with the appropriate construct. Cells are cultured on unlabeled M9 medium in 1H2O, typically overnight to an optical density (OD)600 of less than 0.2, collected by centrifugation, washed in sterile phosphate-buffered saline, and resuspended in labeled minimal media with 2H2O, typically 2 g/L glucose and 1 g/L ammonium chloride. After a short 2- to 3-h adaptation process, cells reach an OD600 of 0.5–0.8. Induction is started with 1 mM isopropyl-β-D-thiogalactopyranoside, and cells are aerated at 37ºC for up to 20 h. Some systems may express better at lower temperatures. 4. Notes 1. If the linker is believed to be entirely uninvolved in the function, then the direct insertion and testing of the construct [Optional Tag 1]-[Domain I with C-terminal (3′) restriction site 1 (RS1)]-IEGRC-[Domain II with N-terminal (5′) restriction site 2 (RS2)]-[Optional Tag 2] for function will then permit a simple PCR process to the intermediary [Optional Tag 1]-[Domain I/RS1]-G-Intein and [Optional Tag III]-IEGR-C-[Domain II/RS2]-[Optional Tag 2]. 2. Ethanethiol presents little hazard to experienced chemists working in well-ventilated hoods. However, users should consult their material data safety sheets, or the local equivalent, before using this flammable, volatile, and odiferous material. 3. It cannot be sufficiently emphasized that the conditions of reaction require highly purified materials of exactly the sequences designed. Therefore, it is essential to use DNA sequencing on all new vectors, and to use mass spectrometry to identify rigorously the unlabeled precursors, before runs with isotope labeling. Measurements of protein concentrations in mixtures can be estimated for gels, or directly measured against standards using analytical HPLC.

References 1. 1 Muller, A., MacCallum, R. M., and Sternberg, M. J. (2002) Structural characterization of the human proteome. Genome Res. 12, 1625–1641. 2. Riek, R., Fiaux, J., Bertelsen, E. B., Horwich, A. L., and Wuthrich, K. (2002) 2 Solution NMR techniques for large molecular and supramolecular structures. J. Am. Chem. Soc. 124, 12,144–12,153. 3. Pervushin, K., Riek, R., Wider, G., and Wuthrich, K. (1997) Attenuated T2 3 relaxation by mutual cancellation of dipole–dipole coupling and chemical shift anisotropy indicates an avenue to NMR structures of very large biological macromolecules in solution. Proc. Natl. Acad. Sci. USA 94, 12,366–12,371. 4. The Grove Concise Dictionary of Music (1994) (Sadie, S., ed.), Macmillan, New York, p. 541.

Segmental Isotopic Labeling

55

5. Ottesen, J. J., Blashke, U. K., Cowburn, D., and Muir, T. W. (2003) Segmental isotopic labeling: prospects for a new tool to study the structure–function relationships in multidomain proteins. In Biological Magnetic Resonance, vol. 20 (Krishna, N. R. and Berliner, L. J., eds.), Kluwer Academic, New York, pp. 35–51. 6. Dawson, P. E., Muir, T. W., Clark-Lewis, I., and Kent, S. B. H. (1994) Synthesis of 6 proteins by native chemical ligation. Science 266, 776–779. 7. 7 Xu, M. Q., Comb, D. G., Paulus, H., Noren, C. J., Shao, Y., and Perler, F. B. (1994) cinimide formation. EMBO J. 13, 5517–5522. 8. Chong, S., Shao, Y., Paulus, H., Benner, J., Perler, F. B., and Xu, M. Q. (1996) Protein splicing involving the Saccharomyces cerevisiae VMA intein: the steps in the splicing pathway, side reactions leading to protein cleavage, and establishment of an in vitro splicing system. J. Biol. Chem. 271, 22,159–22,168. 9. 9 Paulus, H. (2000) Protein splicing and related forms of protein autoprocessing. Annu. Rev. Biochem. 69, 447–496. 10. Liu, X. Q. (2000) Protein-splicing intein: genetic mobility, origin, and evolution. 10 Annu. Rev. Genet. 34, 61–76. 11. 11 Yamazaki, T., Otomo, T., Oda, N., Kyogoku, Y., Uegaki, K., and Ito, N. (1998) Segmental isotope labeling for protein NMR using peptide splicing. J. Am. Chem. Soc. 120 (22), 5591,5592. 12. 12 Otomo, T., Ito, N., Kyogoku, Y., and Yamazaki, T. (1999) NMR observation of selected segments in a larger protein: central-segment isotope labeling through intein-mediated ligation. Biochemistry 38, 16,040–16,044. 13. Otomo, T., Teruya, K., Uegaki, K., Yamazaki, T., and Kyogoku, Y. (1999) Improved 13 segmental isotope labeling of proteins and application to a larger protein. J. Biomol. NMR 14, 105–114. 14. 14 Xu, M. Q. and Evans, T. C., Jr. (2001) Intein-mediated ligation and cyclization of expressed proteins. Methods 24, 257–277. 15. Camarero, J. A., Shekhtman, A., Campbell, E. A., Chlenov, M., Gruber, T. M., 15 Bryant, D. A., et al. (2002) Autoregulation of a bacterial σ factor explored by using segmental isotopic labeling and NMR. Proc. Natl. Acad. Sci. USA 99, 8536–8541. 16. 16 Xu, R., Ayers, B., Cowburn, D., and Muir, T. W. (1999) Chemical ligation of folded recombinant proteins: segmental isotopic labeling of domains for NMR studies. Proc. Natl. Acad. Sci. USA 96, 388–393. 17. 17 Evans, T. C., Jr., Benner, J., and Xu, M. Q. (1998) Semisynthesis of cytotoxic proteins using a modified protein splicing element. Protein Sci. 7, 2256–2264. 18. 18 Erlanson, D. A., Chytil, M., and Verdine, G. L. (1996) The leucine zipper domain controls the orientation of AP-1 in the NFAT.AP-1.DNA complex. Chem. Biol. 3, 981–991. 19. 19 Cowburn, D. and Muir, T. W. (2001) Segmental isotopic labeling using expressed protein ligation. Meth. Enzymol. 339, 41–54. 20. 20 Tolbert, T. J. and Wong, C.-H. (2002) New methods for proteomic research: preparation of proteins with N-terminal cysteines for labeling and conjugation. Angew. Chem. Int. Ed. Engl. 114, 2275–2278.

56

Cowburn et al.

21. 21 Gardner, K. H. and Kay, L. E. (1998) The use of 2H, 13C, 15N multidimensional NMR to study the structure and dynamics of proteins. Annu. Rev. Biophys. Biomol. Struct. 27, 357–406. 22. Shekhtman, A., Ghose, R., Goger, M., and Cowburn, D. (2002) NMR structure 22 determination and investigation using a reduced proton (REDPRO) labeling strategy for proteins. FEBS Lett. 524, 177–182. 23. 23 Leiting, B., Marsilio, F., and O’Connell, J. F. (1998) Predictable deuteration of recombinant proteins expressed in Escherichia coli. Anal. Biochem. 265, 351–355. 24. Marley, J., Lu, M., and Bracken, C. (2001) A method for efficient isotopic labeling of recombinant proteins. J. Biomol. NMR 20, 71–75.

5 TROSY-Based Correlation and NOE Spectroscopy for NMR Structural Studies of Large Proteins Guang Zhu, Youlin Xia, Donghai Lin, and Xiaolian Gao Summary Transverse relaxation-optimized spectroscopy (TROSY) is based on the fact that cross-correlation relaxation rates associated with the interferences between chemical shift anisotropy and dipole–dipole interactions can be dramatically reduced. TROSY selects these slowly relaxing components of 15N–1HN or 13C–1H antiphase coherences to significantly enhanced signal sensitivity and spectral resolution for large proteins (>30 kD). The basic principles and applications of three- and four-dimensional TROSY-based triple-resonance experiments and NOESY experiments for structure-function studies of proteins are discussed in this chapter. To make applications of these experiments easier, some of the experimental setups are also described. Key Words: TROSY; TROSY–HN(CO)CACB; TROSY–HNCACB; TROSY–HNCA; TROSY–HNCO; NOESY–TROSY.

1. Introduction Conventional multidimensional nuclear magnetic resonance (NMR) spectroscopy for structure–function studies of biological macromolecules has been limited to molecules smaller than 50 kDa (1,2). Yet, NMR is one of the principle experimental techniques in structural biology. In recent years, the introduction of transverse relaxation-optimized spectroscopy (TROSY)-based NMR methods (3–8) has opened a new avenue for the structure determination of larger biological macromolecules, particularly for proteins and nucleic acids. As has been demonstrated recently, the cross-correlated relaxation-induced polarization transfer (CRIPT)–TROSY-type experiments (9), modified versions of the first-introduced insensitive nucleus enhancement by polarization transfer (INEPT)–TROSY-type experiments, enable scientists to study proteins up to 900 kDa (10).

From: Methods in Molecular Biology, vol. 278: Protein NMR Techniques Edited by: A. K. Downing © Humana Press Inc., Totowa, NJ

57

58

Zhu et al.

TROSY (3) is based on the fact that cross-correlated relaxation introduced by the interference of the dipole–dipole interaction (DD) and chemical shift anisotropy (CSA) gives rise to much smaller transverse relaxation rates in a high field for a system of two coupled spins (1/2), I and S, such as the 15N-1H moiety in protein backbone (3) and 13C–1H moiety in the aromatic group of amino acids (4), allowing NMR studies of much larger proteins and nucleic acids. For structures of proteins in the molecular weight range up to about 150 kDa, the TROSY scheme is sufficient and effective for obtaining workable correlation spectra (5–8,11,12), triple-resonance spectra for sequential assignment (8,13–16), and nuclear Overhauser effect spectroscopy (NOESY) spectra for resonance assignment and the collection of conformational constraints (11,17–20). It has also been demonstrated that TROSY-based triple-resonance experiments are more sensitive than the conventional triple-resonance experiments for a 23-kDa doubly or triply labeled protein (8). 1.1. TROSY-Based Correlation Spectroscopy In a high magnetic field, both protein backbone 15N-1H and aromatic side chain 13C–1H moieties have a TROSY effect, which is more pronounced if proteins are deuterated (8).The TROSY effect will yield the narrowest line width for 1HN at 1 GHz and for 15N at 900 MHz (21). This property can be readily used in many types of correlation spectroscopy applied to isotopically labeled proteins or nucleic acids. There are some fundamental differences between two-dimensional (2D) 15N1H TROSY and 15N-1H heteronuclear single-quantum coherence (HSQC). 2D TROSY spectra of a deuterated protein manifest much narrower line widths, the peak positions are shifted by 1JHN/2 ≅ 45 Hz in the 15N and 1HN dimensions, and the NH2 proton resonances do not appear as in 2D HSQC spectra. Other characteristics of 2D TROSY are similar to that of 2D HSQC, as are their applications and experimental setups. It is well known that the single transition to single transition polarization transfer TROSY can only select 50% of the spin population to have a TROSY effect. The advantage of TROSY-type experiments can only be effectively realized when the transverse relaxation time, T2, of a protein or nucleic acid is sufficiently short (Table 1), such that the spectral line-narrowing effect of TROSY can compensate for the intrinsic lower sensitivity of TROSY. To achieve this experimentally, a longer evolution time is required. 1.2. 3D TROSY-Based Triple-Resonance Experiments TROSY was originally introduced for the study of large deuterated proteins and protein complexes. Therefore, the backbone assignment strategy is slightly different from that of conventional multidimensional NMR methods. For protein

NMR Structural Studies of Large Proteins

59

Table 1 Transverse 1HN and 15N Relaxation Times Predicted for a 23-kDa Protein (τc = 15 ns) at 750 MHz in TROSY-HNCA and HSQC-HNCAa TROSY-HNCA

HSQC-HNCA

Sensitivity gain

R2(15N)[s−1] R2(1HN)[s−1] R2(15N)[s−1] R2(1HN)[s−1] Isolated 15N-1H group β-Sheet 13C/15N-labeled α-helix 13C/15N-labeled β-Sheet 2H/13C/15N-labeled α-helix 2H/13C/15N-labeled a(Reproduced

3.0

3.2

20.9

20.3

8.0

10.6

41.1

28.5

58.2

1.8

8.7

31.5

26.6

48.6

2.0

3.7

6.3

21.6

23.5

4.7

5.0

13.2

22.6

30.3

2.9

with permission from ref. 8.)

backbone assignments, a series of three-dimensional (3D) and four-dimensional (4D) triple-resonance experiments (Table 2) have been developed such as 3D TROSY–HNCA (8,13,14,16,22,23), TROSY–HN(CO)CA (13,16), TROSY–HNCO (8,13,16,24), TROSY–HN(CA)CO (13,16), TROSY–HNCACB (13,16,25), TROSY–HN(CO)CACB (13,16), 4D-TROSY–HNCACO (26), and TROSY–HNCOCA (26). Other types of correlation spectroscopy, such as [13C, 1H]-TROSY–HNC (4,27–29), have also been proposed. The basic design in TROSY-based triple-resonance experiments is the same as for conventional ones with the only exception being that in the TROSY experiments, one has to ensure that the slowly relaxing magnetization components are retained during evolution of the magnetization throughout the whole pulse sequence and during detection. Similar to the conventional HSQC-based experiments, 3D TROSY–HNCA (HNi, Ni, Cαi, {Cαi–1}) (8,13,14,16,22,23), 3D TROSY–HN(CA)CO (HNi, Ni, C′i) (13,16), and TROSY–HNCACB (HNi, Ni, Cαi, {Cαi–1}, Cβi, {Cβi–1}) (13,16,25) provide intraresidue correlations (Table 2). The smaller 2JNCα coupling can sometimes provides weak interresidue peaks in TROSY–HNCA and TROSY–HNCACB spectra. These experiments rely on the 1JHN ≅ 90 Hz, 1JNCα = 8–12 Hz and 1JCC′ = 55 Hz couplings for the magnetization transfer. For obtaining interresidue correlations, TROSY–HN(CO)CACB (HNi, Ni, α C i–1, Cβi–1) (13,16), TROSY–HN(CO)CA (HNi, Ni, Cαi–1) (13,16), and TROSY–HNCO (HNi, Ni, C′i–1) (8,13,16,24) (Table 2) have been designed with

Table 2 TROSY based Triple Resonance Experiments Used for Backbone Assignment

60

61

62

Zhu et al.

the use of the 1JHN, 1JNC′ (≅ 15 Hz) and 1JCC′ couplings for the magnetization transfer. 1.3. 4D TROSY-Based Triple-Resonance Experiments Because the T2 relaxation time in TROSY is much longer compared with that in HSQC, it becomes ideal to design 4D TROSY-based triple-resonance experiments for the backbone assignment of larger 13C/15N/2H-labeled proteins (30). 4D TROSY–HNCACO (HNi, Ni, Cαi, {Cαi−1}, COi′, {CO′i−1}), TROSY–HNCOCA (HNi, Ni, Cαi−1, CO′i−1), (26) and TROSY–HNCOi−1CAi (HNi, Ni, Cαi, C′i−1 ) (31) provide both inter- and intraresidue correlations based on 1JHN, 1JNCα, 1JNC′ , and 1JCC′ for the magnetization transfer. A list of 3D and 4D triple-resonance experiments used for backbone assignment is shown in Table 2. 1.4. Application of TROSY-Based Experiments for Protein Backbone Assignment (30) The TROSY-based triple-resonance experiments described in the preceding sections have been used for the backbone assignment of 67-kDa 13C/15N/2Hlabeled p53 in a manner similar to that used in HSQC-based multidimensional NMR (30). Spectra of the 3D and 4D experiments were processed with the use of linear prediction methods (32,33) to extend the truncated free induction decays (FIDs) in the indirectly detected dimensions before the use of conventional Fourier transformation (FT) procedures to enhance resolution and reduce truncation artifacts. Protein sample used: 0.4 mM 15N/13C/2H-labeled p53 (393 a.a., dimer, 67 kDa) in 25 mM phosphate, 250 mM NaCl, 10 mM dithiothreitol (DTT), 5 mM MgCl2, pH 6.5, 90% H2O/10% D2O. Equipment: 600 MHz NMR spectrometer with three channels. Temperature: 25ºC. Assignment strategy: If (13Cα, 13C′) and (1HN, 15N) spin pairs are unique, then it is possible to link (1HN, 15N) of the ith residue to that of the i+1th residue through the common (13Cα, 13C′) chemical shifts with the use of the 4D TROSY–HNCACO (HN, Ni, Cαi, CO′i), and TROSY–HNCOCA (HNi, Ni, Cαi−1, CO′i−1) experiments. In the case of chemical-shift degeneracy of (13Cα, 13C′) or (1HN, 15N) spin pairs, additional experiments must be analyzed, such as 4D 15N/15N–NOESY–HSQC or 4D 15N/15N–NOESY–TROSY (20) (see Subheading 2.3.), 4D TROSY– HNCOi−1CAi (HNi, Ni, Cαi, C′i−1) (31), 3D TROSY–HN(CA)CB (1HNi, 15Ni, 13 β C i) and 3D TROSY–HN(COCA)CB (1HNi−1, 15Ni−1, 13Cβi) (13). In addition to the above backbone assignment strategy using 4D TROSYbased experiments, 3D TROSY-based experiments are widely employed to

NMR Structural Studies of Large Proteins

63

obtain backbone assignments of a protein. The inter- and intraresidue connections can be readily obtained using the 3D TROSY-based experiments listed in Table 2. Normally, Cβ chemical shifts have much less degeneracy. Hence, the 3D TROSY–HNCACB (1HNi, 15Ni, 13Cαi {13Cαi−1}, 13Cβi {13Cβi−1}) and 3D TROSY–HN(CO)CACB (1HNi−1, 15Ni−1, 13Cαi, 13Cβi) (13,16,25) and 3D NOESY–TROSY (11) are useful in resolving ambiguous assignments. 2. TROSY-Based NOESY Experiments NMR-derived protein structures are based on interproton distance restraints obtained from 15N- and 13C-edited 2D, 3D, and 4D NOESY data sets, torsion angle restraints generated from various heteronuclear and homonuclear scalar coupling correlation experiments (34,35), and residual dipolar couplings obtained using samples dissolved in liquid crystal media (36–38). Among these, NOE restraints are essential for defining the structure to high resolution. Specifically, NOE restraints are obtained from 3D 15N-edited (39), 13C-edited (40), or 3D or 4D 15N- and 13C-edited NOESY experiments (41,42). TROSY-based NOESY experiments (11,17–20,43–46) have been shown to be superior in signal enhancement for large 15N/13C/2H-labeled proteins. In the following sections, these 15N-separated TROSY-based NOESY experiments (11,17,18,20) are discussed in detail. 2.1. 3D Single-Quantum NOESY–TROSY Experiment 3D single-quantum NOESY–TROSY (11) was the earliest experiment used to introduce TROSY into 3D 15N separated NOESY experiments. In the 3D single-quantum NOESY–TROSY, water flip-back en-TROSY is used so that the sensitivity is enhanced by a factor of 2 compared with the original version of TROSY (3), and the signal attenuation resulting from the chemical exchange between water protons and amide protons is reduced. The pulse sequence of this experiment is shown in Fig. 1. The advantage of TROSY-type experiments can only be effectively realized when the transverse relaxation time, T2, of a protein or nucleic acid is small enough such that TROSY can have better sensitivity than HSQC. To achieve this experimentally, a longer evolution time is required. With this in mind, single-quantum magnetization is used during the t1 evolution time to obtain higher sensitivity as depicted in the Fig. 1. Application to proteins: Protein sample used: 1 mM 15N-labeled calmodulin (147 a.a.) in 90% H2O/10% D2O at pH 6.8. Equipment: 500-MHz NMR spectrometer with three channels. Experimental parameters and results: Temperature: 15ºC. Experimental parameters: WATERGATE uses a 1.7-ms pulse with a power of 110 Hz. τ1 = 2.7 ms and τmix = 100 ms (Fig. 1). The quadrature component

64

Zhu et al.

Fig. 1. Pulse sequence of the 3D single quantum NOESY-TROSY experiment. Filled bars and open bars represent 90º and 180º pulses, respectively. Default phases are x. ϕ = 4(x), 4(−x); ϕ2 = (y, x, −y, −x), ϕ = (y); (x, −y, −x, y, −x, y, x, −y).

in the t1 dimension is acquired by altering the phase of ϕ1 in the States-time proportional phase increment (TPPI) manner. Echo/antiecho selection during t2 is done by reversing ϕ3 and the even-number phases of the receiver. The durations and strengths of the gradients are of conventional values and the bipolar gradient G2 is 0.5 G/cm. Processing parameters: A 3D NOESY–TROSY data matrix of 64×32×512 complex points was acquired. The spectrum (Fig. 2) was processed with the use of linear prediction in the t1 and t2 time domains to extend the size of the FIDs by 50% (33,47). A cosine-square bell window function was applied in t1 and t2, and the data are zero-filled one time before FT. The 3D single quantum NOESY–TROSY experiment is widely used to obtain distance restraints and to verify assignments for doubly and triply labeled large proteins because of its simplicity and sensitivity. This experiment can be applied to triply labeled proteins above approx 40 kDa to provide the following NOEs: dNN(i, i+1), (Strong [α-helix], Weak [β-sheet]), dNN (medium [crossβ strands]) and other long-range NOEs used for obtaining spectral assignments and distance constrains. 2.2. 2D and 3D Transverse Relaxation-Optimized NOESY Experiments When the T2 relaxation time of a protein or nucleic acid is short enough and the t1 evolution time is sufficiently long, the sensitivity of the single-quantum NOESY–TROSY experiment described in the preceding section is no longer optimized. In these cases, it is necessary to enhance the 3D NOESY experiment by TROSY in all three dimensions (17,19,46). Here, two sensitivity-enhanced NOESY–TROSY experiments, namely, the 2D spin-state selective excitation

NMR Structural Studies of Large Proteins

65

method (S3E)–NOESY–S3E and 3D TROSY–NOESY–S3E experiments (17), with the use of the S3E (48,49) are described. These experiments have the TROSY effect in all of the indirectly and directly detected dimensions, and so provide optimal resolution of amide protons. Furthermore, these two experiments provide an additional useful feature in that the diagonal peaks of the amide proton region are canceled or greatly reduced. This enables clear identification of NOESY cross peaks close to the diagonal peaks. The pulse sequences of the 2D S3E–NOESY–S3E and 3D TROSY– NOESY–S3E experiments are depicted in Fig. 3A and B, respectively. The 2D S3E–NOESY–S3E experiment is a combination of the S3E method and the ordinary NOESY experiment. This experiment not only gives HN–HN and HN–HC (for partially deuterated proteins) NOE cross peaks with the characteristic narrow TROSY HN line widths and greatly reduced diagonal peaks, but it also provides normal NOE cross-peaks, although with reduced intensities. In the 3D TROSY–NOESY–S3E experiment, the selection of the slowly relaxing components of 1H and 15N magnetization before the NOE mixing time is achieved by the use of the 15N-1H TROSY scheme, and the selection of the TROSY 1H component before detection is accomplished by the use of the S3E

Fig. 2. 2D cross-section taken form 3D NOESY-TROSY spectrum (at 15N = 127.7ppm). The peak at 1H = 4.8 ppm is the exchange peak between water and NH of A 147 at the terminus of the protein. The NOE mixing time is 100 ms. (Reproduced from ref. 11 with kind permission of Kluwer Academic Publishers.)

66

Zhu et al.

Fig. 3. 2D and 3D TROSY enhanced NOESY pulse sequences. Filled bars and open bars represent 90º and 180º pulses, respectively. Default phases are in the x direction. (A) S3E-NOESY-S3E pulse sequence: ϕ1 = (x, −x, −x, x, y, −y, −y, y) + 45º; ϕ2 = 4(y), 4(−x); ϕ3 = 4(x), 4(y); ϕ4 = (x, −x, x, −x, y, −y, y, −y); ϕ6 = 4(x), 4(y); ϕ7 = [16(x), 16(y), 16(−x), 16(−y)] + 45º; ϕ8 = 16(y), 16(−x); ϕ9 = 16(x), 16(y); ϕ10 = 8(x), 8(−x), 8(y), 8(−y); ϕr = 4(x, x, −x, x), 4(y, −y, −y, y), 4(−x, x, x, −x), 4(−y, y, y, −y). (B) TROSY-NOESY-S3E pulse sequence: ϕ1 = (x, y, −x, −y); ϕ2 = (y); ϕ3 = (x); ϕ4 = (x, y); ϕ5 = [4(x), 4(−x)] + 45º; ϕ6 = 8(x), 8(−x); ϕ10 = (−x, x); ϕr = (x, x, −x, −x, −x, −x, x, x). (Reproduced from ref. 17 with kind permission of Kluwer Academic Publishers.)

sequence. Thus, the characteristic TROSY spectral line widths are obtained in all three dimensions of the 3D spectrum. The diagonal peaks in the HN–HN region in the two 2D S3E–NOESY–S3E and 3D TROSY–NOESY–S3E spectra are canceled or greatly reduced, as discussed in detail in the literature (17,19,46). The drawback of these experiments is that they are less sensitive than the corresponding normal 2D NOESY and 3D NOESY–HSQC experiments. However, this sensitivity loss is compensated by the narrower spectral line widths afforded by TROSY for large proteins (3). Partial deuteration of proteins can largely reduce the contributions of 1H–1H and 1H–15N dipolar interactions to the relaxation of the HN and N, so that the TROSY effect can be enhanced (8). Application to Proteins: Protein sample used: 1 mM 15N-labeled calmodulin (147 a.a.) in 90% H2O/10% D2O at pH 6.8. Equipment: 750-MHz NMR spectrometer with three channels. Experimental parameters and results: Temperature: 15ºC. Experimental parameters: The experimental recovery delay is 1 s, τ = 1/(41JNH) ≈ 2.7 ms, τm = 80 ms. (A) S3E–NOESY–S3E pulse sequence: The quadrature

NMR Structural Studies of Large Proteins

67

component in the t1 dimension is acquired by altering the phases ϕ1 to ϕ4 in the States-TPPI manner. (B) TROSY–NOESY–S3E pulse sequence: Four transients, Sphase1,phase2, are acquired through altering ϕ2 and ϕ4 (if phase1 = 2, ϕ2 = ϕ2 + 180º and ϕ4 = ϕ4 + ϕ10, otherwise ϕ2 and ϕ4 are unchanged; if phase2 = 2, ϕ4 = ϕ4 −90º, otherwise ϕ4 is unchanged) for every pair of t1 and t2 values. The proton transmitter offset before the NOE mixing period is set at 8.5 ppm. The durations and strengths of the gradients are of conventional values. Processing parameters: In the 2D S3E–NOESY–S3E, spectral widths in both dimensions were 10,500 Hz; 600 FIDs were acquired each with 64 transients. In the 3D TROSY–NOESY–S3E, 64(15N) × 128(1H) FIDs were recorded each with 16 scans. A cosine-square bell window function was applied in t1 and t2, and the data were zero-filled one time before FT. Results: Fig. 4A and B show the amide proton regions of a conventional 2D NOESY spectrum and the 2D S3E–NOESY–S3E spectrum, respectively. It is clear that the diagonal peaks from Fig. 4B are canceled or greatly reduced in the HN region of the spectrum, revealing more cross-peaks with enhanced

Fig. 4. 2D and 3D transverse relaxation optimized NOESY spectra. (A) and (B) show the amide proton region of a normal 2D NOESY spectrum and the S3E-NOESYS3E 2D spectrum, respectively, and (C) the 2D spectrum obtained by the projection of 2D 1H-1H cross-section of the 3D TROSY-NOESY-S3E spectrum along the 15N dimension. (Reproduced from ref. 17 with kind permission of Kluwer Academic Publishers.)

68

Zhu et al.

resolution when compared with the normal NOESY spectrum (Fig. 4A). The line-widths in Fig. 4B are about 4 Hz narrower on average than those in Fig. 4A. For deuterated proteins, the spectral line-widths will be further reduced. The open contours near the diagonal peaks in Fig. 4B are the negative 1JHN splittings of residual diagonal peaks. Figure 4C displays the 2D NOESY spectrum obtained by projecting 2D 1H–1H cross-sections of the 3D TROSY–NOESY–S3E spectrum along the 15N dimension. As seen in Fig. 4C, the 3D TROSY–NOESY–S3E experiment can produce a spectrum with canceled or greatly reduced diagonal peaks in the HN region, but with some loss of cross-peaks in the HN–Hα region resulting from the water suppression scheme used in the S3E (17). 2.3. 3D/4D

15N/15N

Separated HSQC–NOESY–TROSY Experiment

3D single quantum NOESY–TROSY (11) and 3D TROSY–NOESY experiments (17,19,46) have been proposed to measure NOEs between protons. Nevertheless, the 3D NOESY–TROSY experiments may still be insufficient to provide unambiguous measurements of NOEs among the overlapping proton signals. This is particularly acute in proteins with a high level of helical content. To further alleviate the spectral overlap problem, four NOESY experiments, namely, 3D and 4D 15N/15N separated HSQC–NOESY–TROSY and TROSY– NOESY–TROSY have been proposed (20). Perdeuteration of the protein sample not only increases the amide proton T2 and TROSY effect dramatically, but it also reduces the occurrence of spin diffusion, permitting detection of NOEs between protons separated more than 5Å, by using longer NOE mixing times (50,51). Hence, the TROSY-based 3D and 4D 15N/15N separated NOESY experiments could be very useful for structure determination of large proteins. The two pulse sequences are shown in Fig. 5A and B, respectively. In these two pulse sequences, the TROSY sequences replace the latter gradient- and sensitivity-enhanced HSQC sections in the 3D and 4D 15N/15N separated HSQC–NOESY–HSQC pulse sequences. The ideas underlying use of HSQC or TROSY before the mixing time are the same as those discussed in Subheading 2.1. Application to Proteins: Protein sample used: 1 mM 100% 15N/ 70% 2H-labeled trichosanthin (256 a.a.) in 90% H2O/10% D2O at pH 7.0. Equipment: 750-MHz NMR spectrometer with three channels. Experimental parameters and results: Temperature: 5ºC. Experimental parameters: Experimental recovery delay is 1 s, τ = 2.5 ms, τm = 100 ms. The durations and strengths of the gradients are of conventional values and gb are smaller bipolar gradients used to suppress the effect of water

NMR Structural Studies of Large Proteins

69

Fig. 5. HSQC-NOESY-TROSY pulse sequences: (A) 3D version and (B) 4D version. Filled bars and open bars represent 90º and 180º pulses, respectively. Fileed shaped pulses are 1.1 ms sinc-modulated rectangular 90º pulses to selectively excite water resonance. (phase1 = 1, 2; phase2 = 1, 2; phase3 = 1, 2), were recorded. Default phases are x. (A): ϕ1 = (x, x, −x, −x); ϕ5 = (y, x, y, x, −y, −x, −y, −x); ϕ6 = (y); ϕr(x, −y, −x, y, −x, y, x, −y). If phase1 = 2, ϕ1 = ϕ1 + 90º, else ϕ1 is not valid ; if phase2 = 2, ϕ6 = ϕ + 180º and the sign of ϕr is reversed for the even number step phases, otherwise ϕ6 and ϕr are not varied. (B): ϕ1 = (y, −y); ϕ2 = (−x); ϕ3 = (y); ϕ4 (y); ϕ5 = (y, y, x, x); ϕ6 = (y); ϕr = (x, −x, −y, y). If phase1 = 2, ϕ4 = ϕ4 − 90º, otherwise ϕ4 is not varied; if phase2 = 2. ϕ2 = ϕ2 + 180º and the sign of g6 is inverted, otherwise ϕ2 and g6 are not varied; if phase3 = 2, ϕ6 = ϕ6 + 180º and the sign of ϕr is reversed for the third and fourth step phases otherwise ϕ6 and ϕr are not varied. (Reproduced from ref. 20 with kind permission of Kluwer Academic Publishers.)

radiation damping. For the 4D experiment, the proton carrier position between the two arrows is set at 7.7 ppm and elsewhere returned to 4.7 ppm. For the 3D experiment, 4 transients, Sphase1,phase2 (phase1 = 1, 2; phase2 = 1, 2 as explained below), are recorded; for 4D experiment, 8 transients, Sphase1,phase2,phase3, are recorded. Processing parameters: The 3D data matrix was composed of 54×108×1024 complex points with spectral widths of 1825×1825×10,500 Hz. The number

70

Zhu et al.

of scans for each transient was eight. Cosine-bell window functions and zerofilling were used in all dimensions. Results: 2D [15N–15N] (F1–F2) slices of the 3D 15N/15N separated HSQC– NOESY–TROSY spectrum of TCS and its corresponding HSQC–NOESY– HSQC spectrum taken at a 1H chemical shift of F3 = 7.92 ppm are shown in Fig. 6A and B, respectively. The number beside each peak of Fig. 6A is the ratio of peak intensity in Fig. 6A over its corresponding peak intensity in Fig. 6B. For all well-isolated peaks examined, the signal sensitivities in 3D HSQC–NOESY–TROSY experiment are enhanced by 18 to 178% compared with that of 3D HSQC–NOESY–HSQC. On average, a sensitivity enhancement of 62% is obtained. The line widths in the HN (F3) and N (F2) dimensions are reduced by 3–55% and 2–45%, respectively. On average, the line widths are decreased by 20 and 18% in the HN (F3) and N (F2) dimensions, respectively. Two 2D [15N–15N] (F1–F2) slices of the 3D 15N/15N separated HSQC–NOESY–TROSY spectrum taken at 1H chemical shift F3 = 7.66 ppm and F3 = 8.13 ppm are shown in Fig. 7A and B, respectively. For those protons with degenerate 1H chemical shift, their NOE cross-peaks cannot be observed in a normal 2D [1H–1H]–NOESY spectrum, or a 3D 15N separated NOESY spectrum. However, their cross-peaks are clearly seen in Fig. 7. It should be noted that the cross-peaks on the corners of dashed boxes, corresponding to different coherence transfer paths, respectively, are present on the same plane owing to the degeneracy of their proton chemical shifts. The experimental data show that signal sensitivity in 4D HSQC–NOESY– TROSY is enhanced by 8% compared with that of 4D HSQC–NOESY–HSQC experiment. If the number of data points in the t3 dimension is doubled (from 32 to 64 complex points), the sensitivity enhancement would be 18%. For a perdeuterated sample, NOEs between amide protons are essential distance constraints, and the 3D and 4D 15N/15N separated HSQC–NOESY–TROSY experiments proposed here are important to obtain NOEs for large proteins. For even larger proteins (e.g., >80 kD), TROSY–NOESY–TROSY experiment is expected to be more useful (20). 2.4. 3D Zero-Quantum NOESY–TROSY and Other NOESY–TROSY Experiments Transverse relaxation of single-quantum coherences of 15N and 1H can be optimized by exploiting the interference between 15N CSA and the 15N–1H DD interaction (SQ–TROSY) (3). In addition, transverse relaxation of zeroquantum coherence of 15N and 1H can also be optimized by exploiting the interference between 15N and 1H CSAs (ZQ–TROSY) (18). Zero-quantum TROSY has been introduced into the 3D NOESY experiment (18). Here, 2D

NMR Structural Studies of Large Proteins

71

Fig. 6. 2D [15N–15N]-slices of a 15N/15N separated 3D HSQC-NOESY-TROSY spectrum (A) and its corresponding HSQC-NOESY-HSQC spectrum (B) of trichosanthin taken at 1H chemical shift F3 = 7.92 ppm. The chemical shifts of the N2 and HN dimension in A are shifted by 45 Hz compared to those in B. The numbers beside peaks of A are the intensity ratios of peaks in A and their corresponding peaks in B. A and B have the identical lowest contour level and a contour factor of 1.2. (Reproduced from ref. 20 with kind permission of Kluwer Academic Publishers.)

Fig. 7. 2D [15N–15N] (F1–F2) slices of the 3D 15N/15N separated HSQC-NOESYTROSY spectrum taken at 1H chemical shifts of F3 = 7.66 ppm (A) and F3 = 8.13 ppm (B), respectively. (Reproduced from ref. 20 with kind permission of Kluwer Academic Publishers.)

72

Zhu et al.

Fig. 8. Schemes for 2D ZQ-TROSY (A) and3D NOESY-ZQ-TROSY (B). Filled bars and open bars represent 90º and 180º pulses, respectively, and curved shapes represent water-selective 90º rf pulses. (A) The phases of the rf pulses are: φ1 = (−x); φ2 = (x); φ3 = (x, −x, −y, y); ψ1 = (−x, x, −y, y); ψ2 = (y, −, x, −x,); ψ3 = (y); x on all other pulses. To obtain a complex interferogram, a second FID is recorded for each t1 delay, with the following phases: φ1 = (x); φ2 = (−x); φ3 = (x, −x, y, −y); ψ3 = (−y). (B) The phases of the rf pulses are: φ1 = (4(45º), (4(225º); φ2 = (X); φ3 = (−x); φ4 = (x); φ5 = (x, −x, −y, y, −x, x, y, −y); ψ1 = (−x, x, −y, y); ψ2 = (y, −y, x, −x); ψ3 = (y); x on all other pulses. Quadrature detection in t1 dimension is achieved by the States-TPPI method applied with phase φ2. For each t2 increment, a second FID is recorded with the following different phases: φ3 = (X); φ4 = (−x); φ5 = (x, −x, y, −y, −x, x, −y, y); ψ3 = (−y). Details of data processing may be found in the literature (18). (Reproduced with permission from ref. 18 copyright [1999] National Academy of Sciences, U.S.A.)

zero-quantum TROSY and 3D zero-quantum NOESY–TROSY experiments are described. The pulse sequences of 2D ZQ–TROSY and 3D NOESY–ZQ–TROSY are shown in Fig. 8. In the 2D ZQ–TROSY experiment, two different types of the TROSY cross-correlation effects are involved. The cross-correlation effect of CSAs of 15N and 1H is used in t1 period, whereas the cross-correlation effect of 1H CSA and 15N–1H DD coupling is used in t2 period. In the 3D NOESY–ZQ–TROSY experiment, the selection of the slowly relaxing magnetization component in the t1 domain is done by S3E, and after the mixing time, ZQ–TROSY is used for detection. Therefore, transverse relaxation is optimized in all three dimensions of 3D NOESY–ZQ–TROSY. Application to Proteins: Protein sample used: 100% 15N- and 70% 2H-labeled protein 7,8-dihydroneopterin aldolase (110 kDa). Equipment: 750-MHz NMR spectrometer with three channels. Experimental parameters and results: Temperature: 30ºC.

NMR Structural Studies of Large Proteins

73

Experimental parameters: τ = 5.4 ms. (A) In 2D ZQ–TROSY, the complex interferogram is multiplied by exp(-iΩHt1) after FT in the ω2 dimension, where ΩH is the offset in the ω2 dimension relative to the 1H carrier frequency in rad.s–1. (B) In 3D NOESY–ZQ–TROSY, the 1H carrier frequency offsets are set to 8.7 ppm at time-point fo1 and to 4.7 ppm at fo2. Results: The ZQ–TROSY experiment is evaluated in Fig. 9. by comparison with the single-quantum TROSY and 2D HSQC. The figure shows cross-sections along both the 15N and 1H chemical shift axes of a well separated cross-peak. The 2D ZQ–TROSY (Fig. 9A) provides better sensitivity and resolution than the conventional 2D HSQC (Fig. 9C), but the single-quantum TROSY experiment clearly gives the best result (Fig. 9B). Figure 10A and B present two sets of strips taken from 3D NOESY–ZQ–TROSY and conventional 3D NOESY–HSQC. Cross peaks in Fig. 10A are sharper, and some of them are stronger than that in Fig. 10B. Furthermore, the absence of strong diagonal peaks in 3D NOESY–ZQ–TROSY helps to resolve NOE peaks close to the diagonal. In addition, the resonances of the 15N–1H2 groups of Gln and Asn residues are not suppressed by the ZQ–TROSY scheme. The suppression of the strong diagonal peaks may be useful in practice.

Fig. 9. Comparison of 2D ZQ-Trosy (A), SQ-TROSY (B), and HSQC (C). The measurement time and the experiment setup were identical for all three spectra. (Reproduced with permission from ref. 18 copyright [1999] National Academy of Sciences, U.S.A.)

74

Zhu et al.

Fig. 10. Comparison of corresponding spectral region in a 3D NOESY-ZQ-TROSY (A) and a conventional 3D NOESY-HSQC (B). Both experiments were carried out under the same experimental conditions. NOE connectivities are shown with thin horizontal and vertical lines. (Reproduced with permission from ref. 18 copyright [1999] National Academy of Sciences, U.S.A.)

In addition to the NOESY–TROSY experiment described above, a TROSY–NOESY experiment has been developed for the measurement NOEs using the TROSY effect on the aromatic 13C–1H moiety (43). A NOESY–[15N,1H]/[13C,1H]–TROSY, which can simultaneously detect NOEs between HN and aromatic protons by taking advantage of TROSY effects of 15N–1H and 13C–1H moieties on the backbone and aromatic sidechains of a protein, has also been designed (44). 3. Conclusions The TROSY-based methods represent a major advance in NMR studies of large proteins, nucleic acids, and their complexes. 2D TROSY provides a 15 N–1H fingerprint that is highly sensitive to the changes in the protein and provides an NMR probe for studies of intermolecular interaction of proteins or protein complexes up to 100 kD. If the CRIPT–TROSY is used, this limit can

NMR Structural Studies of Large Proteins

75

go as high as 900 kD (10). TROSY-based experiments can be readily applied to small membrane proteins reconstituted in lipid micelles or water-soluble detergents (52). The most suitable magnetic field for the TROSY effect for 15N is 900 MHz and for 1HN is 1 GHz (21). As the number of high field magnets increases, many more large proteins and protein complexes will be studied by TROSY-based NMR experiments. References 1. Bax, A. (1994) Multidimensional nuclear magnetic resonance methods for protein studies, Curr. Opin. Struct. Biol. 4, 738–744. 2. 2 Gardner, K. H., Zhang, X. C., Gehring, K, and Kay, L. E. (1998) Solution NMR studies of a 42 KDa Escherichia coli maltose binding protein beta-cyclodextrin complex: chemical shift assignments and analysis. J. Am. Chem. Soc. 120, 11,738–11,748. 3. 3 Pervushin, K., Riek, R., Wider, G., and Wüthrich, K. (1997) Attenuated T-2 relaxation by mutual cancellation of DD coupling and chemical shift anisotropy indicates an avenue to NMR structures of very large biological macromolecules in solution. Proc. Natl. Acad. Sci. USA 94, 12,366–12,371. 4. 4 Pervushin, K., Riek, R., Wider, G., and Wüthrich, K. (1998) Transverse relaxation-optimized spectroscopy (TROSY) for NMR studies of aromatic spin systems in C-13-labeled proteins. J. Am. Chem. Soc. 120, 6394–6400. 5. 5 Pervushin, K., Wider, G., and Wüthrich, K. (1998) Single transition-to-single transition polarization transfer (ST2-PT) in [N-15,H-1]-TROSY. J. Biomol. NMR 12, 345–348. 6. 6 Zhu, G., Kong, X., Yan, X., and Sze, K. (1998) Sensitivity enhancement in transverse relaxation optimized NMR spectroscopy. Angew. Chem. Int. Ed. Engl. 37, 2859–2861. 7. Weigelt, J. (1998) Single scan, sensitivity- and gradient-enhanced TROSY for multidimensional NMR experiments. J. Am. Chem. Soc. 120, 10,778, 10,779. 8. 8 Salzmann, M., Pervushin, K., Wider, G., Senn, H., and Wüthrich, K. (1998) TROSY in triple-resonance experiments: new perspectives for sequential NMR assignment of large proteins. Proc. Natl. Acad. Sci. USA 95, 13,585–13,590. 9. 9 Riek, R., Wider, G., Pervushin, K., and Wüthrich, K. (1999) Polarization transfer by cross-correlated relaxation in solution NMR with very large molecules. Proc. Natl. Acad. Sci. USA 96, 4918–4923. 10. 10 Fiaux, J, Bertelsen, E. B., Horwich, A. L., and Wüthrich, K. (2002) NMR analysis of a 900K GroEL-GroES complex. Nature 418, 207–211. 11. 11 Zhu, G., Kong, X. M., and Sze, K. H. (1999) Gradient and sensitivity enhancement of 2D TROSY with water flip-back, 3D NOESY-TROSY and TOCSY-TROSY experiments. J. Biomol. NMR 13, 77–81. 12. 12 Rance, M., Loria, J. P., and Palmer, A. G. (1999) Sensitivity improvement of transverse relaxation-optimized spectroscopy. J. Magn. Reson. 136, 92–101. 13. 13 Salzmann, M., Wider, G., Pervushin, K., Senn, H., and Wüthrich, K. (1999) TROSY-type triple-resonance experiments for sequential NMR assignments of large proteins. J. Am. Chem. Soc. 121, 844–848.

76

Zhu et al.

14. 14 Salzmann, M., Pervushin, K., Wider, G., Senn, H., and Wüthrich, K. (1999) [C13]-constant-time [N-15,H-1]-TROSY–HNCA for sequential assignments of large proteins. J. Biomol. NMR 14, 85–88. 15. Yang, D. W. and Kay, L. E. (1999) Improved lineshape and sensitivity in the 15 HNCO-family of triple resonance experiments. J. Biomol. NMR 14, 273–276. 16. 16 Loria, J. P., Rance, M., and Palmer, A. G. (1999) Transverse-relaxation-optimized (TROSY) gradient-enhanced triple-resonance NMR spectroscopy. J. Magn. Reson. 141, 180–184. 17. 17 Zhu, G., Xia, Y., Sze, K., and Yan, X. (1999) 2D and 3D TROSY–enhanced NOESY of N-15 labeled proteins. J. Biomol. NMR 14, 377–381. 18. 18 Pervushin, K. V., Wider, G., Riek, R., and Wüthrich, K. (1999) The 3D NOESY[H-1,N-15,H-1]-ZQ-TROSY NMR experiment with diagonal peak suppression. Proc. Natl. Acad. Sci. USA 96, 9607–9612. 19. 19 Meissner, A. and Sørensen, O. W. (1999) Suppression of diagonal peaks in TROSY-type H-1 NMR NOESY spectra of N-15-labeled proteins. J. Magn. Reson. 140, 499–503. 20. 20 Xia, Y., Sze, K., and Zhu, G. (2000) Transverse relaxation optimized 3D and 4D N15/N-15 separated NOESY experiments of N-15 labeled proteins. J. Biomol. NMR 18, 261–268. 21. Wüthrich, K. (1998) The second decade-into the third millennium. Nat. Struct. Biol. 5(Suppl. S), 492–495. 22. 22 Pervushin, K., Gallius, V., and Ritter, C. (2001) Improved TROSY–HNCA experiment with suppression of conformational exchange induced relaxation. J. Biomol. NMR 21, 161–166. 23. 23 Permi, P. and Annila, A. (2001) A new approach for obtaining sequential assignment of large proteins. J. Biomol. NMR 20, 127–133. 24. Xia, Y., Sze, K. H., Li, N., Shaw, P. C., and Zhu, G. (2002) Protein dynamics measurements by 3D HNCO based NMR experiments. Spectrosc-Int J. 16, 1–13. 25. Meissner, A. and Sørensen, O. W. (2001) Sequential HNCACB and CBCANH 25 protein NMR pulse sequences. J. Magn. Reson. 151, 328–331. 26. 26 Yang, D. W. and Kay, L. E. (1999) TROSY triple-resonance four-dimensional NMR spectroscopy of a 46 ns tumbling protein. J. Am. Chem. Soc. 121, 2571–2575. 27. 27 Riek, R., Pervushin, K., Fernandez, C., Kainosho, M., and Wüthrich, K. (2001) [C-13,C-13]- and [C-13,H-1]-TROSY in a triple resonance experiment for ribose-base and intrabase correlations in nucleic acids. J. Am. Chem. Soc. 123, 658–664. 28. 28 Meissner, A. and Sørensen, O. W. (1999) Optimization of three-dimensional TROSY-type HCCHNMR correlation of aromatic H-1-C-13 groups in proteins. J. Magn. Reson. 140, 447–450. 29. 29 Simon, B., Zanier, K., and Sattler, M. (2001) A TROSY relayed HCCH-COSY experiment for correlating adenine H2/H8 resoances in uniformly 13C-labeled RNA molecules. J. Biomol. NMR 20, 173–176. 30. 30 Mulder, F. A. A., Ayed, A., Yang, D. W., Arrowsmith, C. H., and Kay, L. E. (2000) Assignment of H-1(N), N-15, C-13(alpha), (CO)-C-13 and C-13(beta) resonances

NMR Structural Studies of Large Proteins

31. 31 32. 33. 34. 34 35. 35 36. 36 37. 37

38. 38 39. 39 40. 41. 41 42. 43. 43 44. 44

45. 45

77

in a 67 kDa p53 dimer using 4D-TROSY NMR spectroscopy. J. Biomol. NMR 18, 173–176. Konrat, R., Yang, D. W., and Kay, L. E. (1999) A 4D TROSY-based pulse scheme for correlating (HNi)-H-1,Ni-15,C-13(i)alpha,C-13’(i-1) chemical shifts in high molecular weight, N-15,C-13, H-2 labeled proteins. J. Biomol. NMR 15, 309–313. Zhu, G. and Bax, A. (1990) Improved linear prediction for truncated signals of known phase. J. Magn. Reson. 90, 405–410. Zhu, G. and Bax, A. (1992) Improved linear prediction of truncated damped sinusoids using modified backward-forward linear prediction. J. Magn. Reson. 100, 202–207. Wagner, G. (1989) Heteronuclear nuclear magnetic-resonance experiments for studies of protein conformation. Methods Enzymol. 176, 93–113. Clore, G. M. and Gronenborn, A. M. (1991) Structures of larger proteins in solution— 3-dimensional and 4-dimensional heteronuclear NMR-spectroscopy. Science 252, 1390–1399. Clore, G. M., Gronenborn, A. M., and Bax, A. (1998) A robust method for determining the magnitude of the fully asymmetric alignment tensor of oriented macromolecules in the absence of structural information. J. Magn. Reson. 133, 216–221. Tjandra, N., Grzesiek, S., and Bax, A. (1996) Magnetic field dependence of nitrogen-proton J splittings in N-15-enriched human ubiquitin resulting from relaxation interference and residual dipolar coupling. J. Am. Chem. Soc. 118, 6264–6272. Tjandra, N. and Bax, A. (1997) Direct measurement of distances and angles in biomolecules by NMR in a dilute liquid crystalline medium. Science 278, 1111–1114. Zuiderweg, E. R. P. and Fesik, S. W. (1989) Heteronuclear 3-dimensional NMRspectroscopy of the inflammatory protein c5a. Biochemistry 28, 2387–2391. Muhandiram, D. R., Farrow, N., Xu, G. Y., Smallcombe, S. H., and Kay, L. E. (1993) A gradient C-13 NOESY-HSQC experiment for recording NOESY spectra of C-13-labeled proteins dissolved in H2O. J. Magn. Reson. B102, 317–321. Kay, L. E., Clore, G. M., Bax, A., and Gronenborn, A. M. (1990) 4-dimensional heteronuclear triple-resonance NMR-spectroscopy of interleukin-1-beta in solution. Science 249, 411–414. Farmer, B. T., II. (1991) Simultaneous [C-13, N-15]-HMQC, a pseudo-tripleresonance experiment. J. Magn. Reson. 93, 635–641. Brutscher, B., Boisbouvier, J., Pardi, A., Marion, D., and Simorre, J. P. (1998) Improved sensitivity and resolution in H-1-C-13 NMR experiments of RNA. J. Am. Chem. Soc. 120, 11,845–13,590. Pervushin, K., Braun, D., Fernández, C., and Wüthrich, K. (2000) [N-15,H-1]/ [C-13,H-1]-TROSY for simultaneous detection of backbone N-15-H-1, aromatic C-13-H-1 and side-chain N-15-H-1(2) correlations in large proteins. J. Biomol. NMR 17, 195–202. Brutscher, B., Boisbouvier, J., Kupce, E., Tisne, C., Dardel, F., Marion, D., Simorre, J.P. (2001) Base-type-selective high-resolution C-13 edited NOESY for sequential assignment of large RNAs. J. Biomol. NMR. 19, 141–151.

78

Zhu et al.

46. 46 Meissner, A. and Sørensen, O. W. (2002) Enhanced diagonal peak suppression in three-dimensional TROSY-type N-15-resolved H-1(N)-H-1(N) NOESY spectra. Concepts Magn. Reson. 14, 1–8. 47. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G, Pfeifer, J., and Bax, A. NMRPIPE 47 —a multidimensional spectral processing system based on Unix pipes. J. Biomol. NMR 6, 277–293. 48. 48 Meissner, A. Duus, J. Ø., and Sørensen, O. W. (1997) Integration of spin-stateselective excitation into 2D NMR correlation experiments with heteronuclear ZQ/2Q pi rotations for (1)J(XH)-resolved E.COSY-type measurement of heteronuclear coupling constants in proteins. J. Biomol. NMR 10, 89–94. 49. Sørensen, M. D., Meissner, A., and Sørensen, O. W. (1997) Spin-state-selective coherence transfer via intermediate states of two-spin coherence in IS spin systems: application to E.COSY-type measurement of J coupling constants. J. Biomol. NMR 10, 181–186. 50. Torchia, D. A., Sparks, S. W., and Bax, A. (1988) Delineation of alpha-helical domains in deuteriated staphylococcal nuclease by 2D NOE NMR-spectroscopy. J. Am. Chem. Soc. 110, 2320–2321. 51. Venters, R. A., Metzler, W. J., Spicer, L. D., Mueller, L., and Farmer, B. T., II. (1995) Use of H1(N)-H1(N) NOEs to determine protein global folds in perdeuterated proteins. J. Am. Chem. Soc. 117, 9592–9595. 52. Fernández, C., Adeishvili, K., and Wüthrich, K. (2001) Transverse relaxationoptimized NMR spectroscopy with the outer membrane protein Ompx in dihezanoyl phosphatidylcholine micelles. Proc. Natl. Acad. Sci. USA 98, 2358–2363.

6 Media for Studies of Partially Aligned States Kieran Fleming and Stephen Matthews Summary Measurement of residual dipolar couplings for proteins in nuclear magnetic resonance (NMR) requires a degree of molecular alignment. This may be achieved through the use of liquid crystals or compressed hydrated gels. Several media have been described in the literature, and this chapter describes five of the most commonly used systems. For two of these systems, bicelles and filamentous bacteriophage, the media can be purchased in a form ready for experiments. The remainder must be made by the investigator, yet they are generally straightforward to synthesize in the laboratory. Poly(ethylene glycol)/alcohol mixtures and Helfrich phases are made by simply mixing the ingredients in the correct manner, and strained gels use techniques familiar to all molecular biologists. Key Words: Liquid crystal; NMR; residual dipolar couplings; bacteriophage; strain-induced alignment; bicelles.

1. Introduction Residual dipolar couplings (RDCs) may be used as restraints in macromolecular structure determinations by nuclear magnetic resonance (NMR). The ability to measure RDC values accurately has resulted in a significant improvement in the accuracy of protein solution structures over the past 2–3 yr, particularly in cases where the molecule under study is multidomain. To measure RDCs, a small degree of alignment of the protein of interest in solution is required. As such, various liquid crystal technologies have been described that confer the necessary partial alignment and retain many of the advantages of solution NMR. The various liquid crystalline media are quite diverse (1–6), and each has its own set of characteristics that makes it suitable for particular applications.

From: Methods in Molecular Biology, vol. 278: Protein NMR Techniques Edited by: A. K. Downing © Humana Press Inc., Totowa, NJ

79

80

Fleming and Matthews

The suitability of a particular liquid crystal system for use in NMR studies is based on several criteria. Essentially, they should: 1. Form a homogeneous liquid crystalline phase at a concentration commensurate with potential NMR studies, that is, at a suitable pH and temperature. 2. Show no deleterious interaction with the molecule under study, that is, they may align the molecule under study via steric and/or electrostatic interactions but should not bind strongly to it. 3. Produce a tunable alignment in the magnetic field, that is, the degree of alignment of a macromolecule can be adjusted by altering the liquid crystal concentration or in the case of gels by the amount of applied strain. 4. Remain stable over a long period of time.

As a result of the rather unpredictable behavior of proteins in anisotropic phases, it is generally desirable to test as many systems as possible to maximize the likelihood of success and to provide more than one alignment tensor. Despite this, the choice of system can be refined based on properties of the protein under study. Notably, in cases of highly charged macromolecules (i.e., proteins with extreme pIs), a neutral system is more likely to prove successful— for example, poly(ethylene glycol)/alcohol or strained gels. Furthermore, for weakly charged proteins, it can be advantageous to choose a system that is oppositely charged (e.g., dimyristoylphosphatidylcholine [DMPC]:dihexanoylphosphatidylcholine [DHPC]:cetyltrimethylammonium bromide [CTAB] for negatively charge proteins). If the target protein is particularly valuable, then a liquid crystal system that is readily removable is preferable—for example, phage or strained gels. Several of the more commonly used media are now commercially available in a form ready to use for NMR experiments; therefore, attempting several systems is not time-consuming. In cases where the medium is not commercially available, details of its manufacture will be provided. However, the preparation of protein-containing samples for NMR is the main focus of this text. 2. Materials One particular point of note when manufacturing liquid crystalline phases is that use of the highest purity materials often results in more stable phases and increases the chances of first-time success. 2.1. Bicelles 1. Premixed lipids for bicelle formation at the desired DMPC:DHPC ratio (Avanti Polar Lipids, Alabaster, AL). These should be stored at −18ºC (see Note 1). 2. Buffer solution: 10 mM sodium phosphate, pH 6.6, 0.15 mM sodium azide, 90% H2O (HPLC-grade), 10% D2O (99.9%). 3. Tetradecyltrimethylammonium bromide (TTAB), 99% (Sigma, Gillingham, UK).

Media for Studies of Partially Aligned States

81

2.2. Filamentous Bacteriophage 1. Pf1 filamentous bacteriophage in 10 mM potassium phosphate buffer, pH 7.6, 2 mM MgCl2, and 0.05% NaN3 (ASLA Biotech, Riga, Latvia).

2.3. Poly(ethylene glycol)/Alcohol Mixtures 1. C12E5—N-dodecyl-penta(ethylene glycol), 98% (Fluka, Buchs, Switzerland). 2. N-hexanol.

2.4. Helfrich Phases 1. 2. 3. 4.

Cetylpyridinium bromide (CPBr) monohydrate, 98% (Sigma, Gillingham, UK). Sodium bromide (NaBr), 99.99+% (Sigma, Gillingham, UK). Hexanol. Relevant NMR buffer, for example, 50 mM (to be diluted to 10 mM) sodium phosphate, pH 5.7, 90% H2O:10% D2O.

2.5. Strained Gels 1. 2. 3. 4. 5.

Acrylamide. N,N'-methylenebisacrylamide. Ammonium persulphate. N,N,N'-tetramethylethylenediamine. Glass mold made from an NMR tube, internal diameter 3.4 mm (524PP, Wilmad Glass). 6. Shigemi microcell NMR tube, internal diameter 4.2 mm (BMS-005B, Shigemi, Allison Park, PA). 7. Buffer solution: 10 mM sodium phosphate, pH 5.7, 90% H2O:10% D2O.

3. Methods 3.1. Bicelles The most commonly used bicelle medium is composed of a mixture of DMPC and DHPC (1,7–12). Mixing of the individual lipids with one another is not recommended, especially as premixed lipid powders for bicelle formation are available for purchase. The example given in this section is for a basic 3:1 DMPC:DHPC suspension. An ideal final concentration for the bicelle suspension for use in protein NMR is around the 3.5% (w/v) to 5% (w/v) mark. 1. Make up a 15% (w/v) stock solution of DMPC/DHPC containing 2.4 mM TTAB (which increases bicelle stability) by adding 50 mg of the premixed lipids to 280 µg of buffer, for example (see Note 2). 2. Vortex for 10 min at 4ºC and leave at this temperature for an additional 15 min. A milky solution should result, where the lipids are not completely dissolved (see Note 3). 3. Vortex again at room temperature for 1 min, and place in a water bath at 38ºC for 30 min.

82

Fleming and Matthews

4. Store at 4ºC for 15 min to allow the mixture to cool, and then vortex at 4ºC for 30 min. 5. Allow to warm to room temperature for 15 min, vortex for 1 min, and then place in a water bath at 38ºC for 30 min. Repeat steps 1–5 until the solution becomes clear and is fluid at low temperatures. At about 25ºC, it should become white and milky and then return to a clear and transparent fluid at 38ºC. The viscosity should also increase markedly at elevated temperatures. The solution may appear to have a slight blue taint (see Note 4). 6. Dilute liquid crystalline NMR samples can be obtained by adding buffer to the 15% (w/v) stock solution followed by vortexing and slow-speed centrifugation. A concentration of 3.5% (w/v) produces a HDO quadrupolar splitting of about 6 Hz at 25ºC (see Note 5). 7. Addition of the diluted solution to a minimal sample of protein, for example, 50 µL, in the same buffer, followed by slow-speed centrifugation completes the preparation (see Note 6).

3.2. Phage Several different bacteriophages have been used for RDC measurement, including Pf1 and fd (2,3,9–11,13–15). Production of this phage can be a tricky process and is not recommended in laboratories where protein expression is regularly undertaken because the risk of cross-contamination is high. Fortunately, Pf1 may be purchased in various buffers ready for use. The following assumes a sample of Pf1 phage in 10 mM potassium phosphate buffer, pH 7.6, 2 mM MgCl2, and 0.05% NaN3 has been purchased (see Note 7). 1. Reduce the phage concentration from 50 mg/mL, as supplied by Asla, to the desired concentration (for example, 15 mg/mL) by weighing out the correct mass of stock solution and adding to buffer. Verification of the concentration can be made by measuring A270 and assuming an extinction coefficient, ε = 2.25 cm–1 mg–1 mL (see Note 8). 2. Centrifuge at 10,000g for 2 min to remove air bubbles resulting from the pipetting, and check the HDO quadrupolar deuterium splitting in the NMR magnet. A concentration of 15 mg/mL should give a splitting of approx 20 Hz at 25ºC. 3. For preparation of protein samples, pipet the phage to the desired concentration into a small volume (50 µL) of protein in the same buffer as the phage. To ensure complete mixing, continuous pipetting followed by brief centrifugation is required (see Note 9). 4. The phage buffer may be changed by spinning down the Pf1 at 370,000g in a Beckman TLA-100.3 rotor for 1 h at 5ºC. Then resuspend the phage in the new buffer. The amount of phage recovered will be only approx 50% of the original yield (http://www.asla-biotech.com/asla-phage.htm).

3.3. Poly(ethylene glycol)/Alcohol Mixtures Although this medium is not commercially available, it is very simple to produce with easily obtainable materials. It is more versatile than bicelles or phage

Partially Aligned States

83

in terms of its stability over a wide pH range. Numerous different poly(ethylene glycol) derivatives may be used, and a choice of alcohol is available in the form of hexanol or octanol (4,9,11,12,15). As with bicelles, use of different proportions of the starting materials yields phases with slightly different properties. In this case, the temperature range over which the phase is stable is dictated by the proportions of the mixed ingredients. One of the basic mixtures, 3% C12E5/hexanol (molar ratio = 0.96), is described. This ordered phase is stable in the temperature range 26–39ºC and, therefore, is suitable for most proteins. 1. Weigh out C12E5 to yield a final concentration of 3% (w/w) and make up to 5 mL with a 90% H2O:10% D2O solution. 2. The pH may be adjusted from the base level of pH 3.5, as desired, using sodium hydroxide. 3. Add the correct weight of hexanol dropwise, while vigorously shaking, to a final molar ratio C12E5:hexanol of 0.96 (see Note 10). 4. Remove air bubbles via centrifugation at 5000g for 2–3 min. 5. Check the HDO quadrupolar deuterium splitting in the NMR spectrometer. A splitting of approx 20 Hz at 38ºC is observed. 6. Pipet 200 µL of the C12E5:hexanol stock solution into a small volume (50 µL) of protein in buffer.

3.4. Helfrich Phases Helfrich phases, the common name for mixtures of CPBr and hexanol, are straightforward to prepare, although care must be taken at several stages during the mixing (5,9,11,12,16). The preparation of a 6.5% (w/v) stock solution is described in the following steps. 1. Weigh out amounts of CPBr and hexanol to achieve a final ratio of 1:1.33 (w/w) in a 6.5% (w/v) solution, and add to the correct volume of a 90% H2O:10% D2O solution. 2. Seal the tube containing the solution, vortex, and heat to 70ºC in a water bath until the solution becomes clear. 3. Leave to cool at room temperature and then centrifuge at 4000g for 2 min. 4. The degree of liquid crystal formation of this stock solution can be tested by leaving the sample to equilibrate in the NMR magnet for at least 6 h and then measuring the HDO quadrupolar deuterium splitting. A splitting of around 20 Hz should be observed at 25ºC (see Note 11). 5. Prepare an NMR buffer in the normal manner, for example, a 50 mM sodium phosphate buffer, pH 5.7, 90% H2O:10% D2O. Also prepare a stock solution of 1 M NaBr (see Note 12). 6. Add the buffer and NaBr solution (to a final concentration of 30 mM) to a sample of protein. Make up this to the volume of an NMR sample—for example, 250 µL—using the stock Helfrich phase and NaBr to a final liquid crystal concentration of 5% (w/v); with the buffer at the desired concentration—for example, 10 mM sodium phosphate and an NaBr concentration of 30 mM (see Note 13).

84

Fleming and Matthews

7. The tube should then be sealed, gentle mixing applied, and centrifuged at 4000g for 2 min. If an excess of hexanol exists, observed as a thin layer on top of the liquid crystal phase, this may be removed carefully using a pipet (see Note 14).

3.5. Strained Gels In addition to liquid crystal media, an alternative approach has recently been adopted whereby the application of compression or stretching forces to a hydrated gel induces the alignment required for residual dipolar coupling measurement (6,9,11,17–20). Termed strain-induced alignment in a gel, this method is particularly attractive to the molecular biologist because it relies on the casting of polyacrylamide gels. 1. Prepare a polyacrylamide gel solution containing 7% (w/v) acrylamide and 0.4% (w/v) N,N'-methylenebisacrylamide. 2. Initiate polymerization via addition of 0.08% (w/v) ammonium persulphate and 0.08% (v/v) N,N,N'-tetramethylethylenediamine. 3. Pipet approx 2.5 mL into the glass mold immediately and leave the gel to set (see Note 15). 4. Once set, the gel should be gently removed from the mold and thoroughly washed with distilled H2O. Submerge the gel in a large excess of distilled H2O under gentle agitation for at least 3–4 h (see Note 16). 5. Cut the gel into sections of the desired length, for example, 25 mm. 6. Introduction of the protein into the gel may be performed in two ways. The success of either method will depend on the protein under study. The first approach is to soak the section of gel in a 250 µL buffered solution containing 2 mM protein and 20% D2O at the desired pH overnight in the Shigemi microcell NMR tube. Typically, the final concentration in the tube can be expected to be around half the starting concentration, with a similar uptake of D2O. The second method is to first soak the gel section in a buffer solution for several hours. Then soak in a small volume, approx 500 µL, of protein solution at a concentration of approx 750 µM overnight. The final concentration in the gel may consequently be up to two-thirds of this value, that is, 500 µM. 7. Once the protein has been added, the plunger of the microcell can be used to apply compression. Typically, a compression ratio (length of gel after compression divided by length of gel before compression) of 0.75 is a good starting point (see Note 17). 8. Use parafilm and tape to hold the plunger in place. 9. A HDO quadrupolar deuterium splitting of a few Hz should be observed at 25ºC.

4. Notes 1. The diacyl phospholipids of regular bicelles can be replaced by their hydrolysisresistant dialkyl analogs, in which the carbohydrate chains are connected to the glycerol part by ether instead of ester bonds—for example, ditetradecyl-PC

Partially Aligned States

2.

3. 4. 5.

6.

7. 8. 9. 10.

11.

85

and dihexyl-PC instead of DMPC and DHPC, respectively. These form liquid crystals over a wide pH range, remain chemically stable for several weeks, and provide the same degree of alignment for macromolecules as the DMPC/DHPC system (21). Because of the instability of bicelle mixtures, it is recommended that solutions are made immediately prior to any experiments. Use of HPLC-grade water is very important in bicelle preparation because the phase is so delicate. DHPC is extremely hygroscopic; prepare solutions in a dry box or dilute with buffer immediately after opening. Use low salt concentration (